pith. sign in

arxiv: 2606.03965 · v1 · pith:O7SIQDSOnew · submitted 2026-06-02 · 💻 cs.CL · cs.AI

Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning

Yu Xia , Zhouhang Xie , Xin Xu , Byungkyu Kang , Prarit Lamba , Xiang Gao , Julian McAuley This is my paper

Pith reviewed 2026-06-28 10:32 UTC · model grok-4.3

classification 💻 cs.CL cs.AI
keywords agentic steeringchain-of-thoughtefficient reasoningLLM controlcontroller agentMarkov decision processreinforcement learningtoken efficiency
0
0 comments X

The pith

A controller agent steers a frozen LLM reasoner through adaptive strategy and phrase actions to match full chain-of-thought accuracy at lower token cost.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Large language models gain accuracy from long chain-of-thought traces but often spend tokens inefficiently and lack inference-time control. The paper treats steering as a Markov decision process in which a separate controller watches the current trace and remaining budget, then chooses a reasoning strategy and a steering phrase to start the next step from the frozen reasoner. The controller begins with synthetic multi-budget trajectories and is refined by reinforcement learning that shapes rewards around the budget. Experiments on multiple benchmarks show the approach reaches full-thinking accuracy while saving tokens and letting users set explicit accuracy-efficiency points. This matters because it keeps the base model unchanged yet adds budget-aware control at inference time.

Core claim

Agentic Chain-of-Thought Steering formulates reasoning control as a Markov decision process where a controller agent, at each step, observes the reasoning trace and remaining thinking budget and issues a steering action that combines a reasoning strategy with a steering phrase; the phrase initiates the next generation step from the frozen reasoner. The controller is initialized on constructed synthetic steering trajectories with multi-budget augmentation and optimized via reinforcement learning that uses budget-conditioned reward shaping, enabling budget-aware strategy selection while preserving the reasoner's generation continuity.

What carries the argument

The controller agent that, inside a Markov decision process, selects a reasoning strategy and steering phrase from the observed trace and remaining budget to direct the frozen reasoner.

If this is right

  • The method reaches full chain-of-thought accuracy while using substantially fewer tokens.
  • Users can set explicit accuracy-efficiency trade-offs at inference time.
  • The approach works across multiple reasoners and tasks without retraining the base model.
  • Generation continuity is maintained because the reasoner itself is never altered.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same controller design could be tested on tasks outside mathematical reasoning such as code generation or multi-step planning.
  • If the synthetic trajectory construction proves robust, it may reduce reliance on human-annotated steering data for future controller training.
  • The budget-conditioned reward shaping might be adapted to other constraints such as latency or memory limits instead of token count.

Load-bearing premise

The controller trained on synthetic steering trajectories with multi-budget augmentation and reinforcement learning will generalize to real benchmarks while preserving the frozen reasoner's continuity and without introducing new errors.

What would settle it

Run the method on the same benchmarks with the controller frozen after training and measure whether final-answer accuracy drops below the full chain-of-thought baseline or token use exceeds the reported savings.

Figures

Figures reproduced from arXiv: 2606.03965 by Byungkyu Kang, Julian McAuley, Prarit Lamba, Xiang Gao, Xin Xu, Yu Xia, Zhouhang Xie.

Figure 1
Figure 1. Figure 1: Overview of ACTS. Left: a controller agent steers a frozen reasoner step by step under a thinking-token budget (Detailed formulation in Section 3.1). Right: an illustrative example of controller-steered reasoner generation. tation to expose the controller to varying budget regimes, and train it with supervised fine-tuning. We then optimize the controller via reinforcement learning with budget-conditioned r… view at source ↗
Figure 3
Figure 3. Figure 3: Budget-conditioned reward shaping. budget fraction bt over the corpus. Although the expert reasoner produces each trace without any budget conditioning, mapping token positions onto our synthetic budget axis exposes a clear temporal structure: UNDERSTAND and PLAN concentrate at the trace opening, EXECUTE holds a broad middle band, CHECK rises through the mid-to-late range, and SUMMARIZE and CONCLUDE domina… view at source ↗
Figure 4
Figure 4. Figure 4: Accuracy vs. Total Tokens across three reasoners (rows) and five benchmarks (columns) under ACTS [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Inference throughput (#Tok/s) comparisons. 5.5 Async controller-reasoner inference incurs negligible latency overhead One practical deployment concern of our ACTS framework is inference latency, since the controller￾reasoner architecture introduces additional con￾troller calls on top of reasoner generation. To mea￾sure how our asynchronous two-server pipeline in Section 3.4 addresses this concern, we bench… view at source ↗
Figure 7
Figure 7. Figure 7: Prompt used by the annotator to jointly classify each reasoning step into one of seven strategies and extract [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: An example of assembled steering trajectory (right) constructed from an annotated reasoning trace (left). [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: System prompt for the controller agent. Controller User Message: First Turn Question: {question} Budget Remaining: 100% Controller User Message: Subsequent Turns Reasoner's Step: {current_step} Budget Remaining: {b_t}% [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Controller user message templates. The first turn supplies the question with full budget; subsequent turns [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Reasoner prompt construction at step t. The reasoner inherits the model’s native chat template; only the thinking trace and steering phrase generated by the controller agent are appended after <think>. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Rescue. Vanilla finds 11,111,111,100 early but then commits to 10,111,111,100 after miscounting its digit sum as 9 (it is 8). ACTS reaches the right candidate via structured stepping and concludes correctly. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Shorten. Vanilla derives 4343 early then re-derives it via powers of 6 and a binary intermediate. ACTS does the division once, verifies once, and concludes. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_13.png] view at source ↗
read the original abstract

Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control. Existing efficient reasoning methods control thinking length by shortening, early-stopping, or compressing traces, leaving how the model thinks implicit. In this paper, we propose Agentic Chain-of-Thought Steering (ACTS), which formulates reasoning steering as a Markov decision process where a controller agent adaptively steers a frozen reasoner during inference. At each step, the controller observes the reasoning trace and remaining thinking budget, then issues a steering action consisting of a reasoning strategy and a steering phrase that initiates the next reasoner step. This enables budget-aware strategy control for efficient reasoning while preserving the reasoner's generation continuity. We initialize the controller agent from our constructed synthetic steering trajectories with multi-budget augmentation, and further optimize it via reinforcement learning with budget-conditioned reward shaping. Experiments across multiple benchmarks show that ACTS matches full-thinking performance with substantial token savings, and enables controllable accuracy-efficiency trade-offs across different reasoners and tasks. The code is available at https://github.com/Andree-9/ACTS.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper proposes Agentic Chain-of-Thought Steering (ACTS), which casts reasoning steering as an MDP in which a controller agent observes the current trace and remaining budget, then emits a strategy-plus-phrase action to steer a frozen reasoner while preserving generation continuity. The controller is initialized on synthetic steering trajectories constructed with multi-budget augmentation and is further optimized by reinforcement learning that uses budget-conditioned reward shaping. The central empirical claim is that, across multiple benchmarks, ACTS matches the accuracy of unrestricted chain-of-thought while delivering substantial token savings and enabling explicit accuracy-efficiency trade-offs for different reasoners and tasks. Code is released at the cited GitHub repository.

Significance. If the transfer from synthetic trajectories to held-out benchmarks can be shown to preserve the frozen reasoner’s behavior without introducing new errors, the method would supply a practical, inference-time mechanism for budget-aware control that does not require retraining the base model. The explicit release of code strengthens reproducibility and allows direct inspection of the synthetic-data pipeline and reward function.

major comments (2)
  1. [Experiments] Experiments section (and abstract): the headline claim that ACTS “matches full-thinking performance with substantial token savings” rests on the unverified assumption that an RL-tuned controller initialized on synthetic multi-budget trajectories will generalize to real benchmarks without injecting distribution shifts or new errors into the frozen reasoner’s generation. No train/test split of benchmarks, no ablation isolating the RL stage, and no quantitative measure of generation continuity (e.g., token-level divergence or error-injection rate) are supplied, rendering the central empirical result impossible to assess.
  2. [Method] Method section (RL optimization paragraph): the budget-conditioned reward shaping is described only at a high level; without the precise functional form of the reward or the synthetic-data construction procedure, it is impossible to determine whether the learned policy is merely memorizing the augmentation distribution rather than learning transferable steering behavior.
minor comments (1)
  1. [Abstract] The abstract states “experiments across multiple benchmarks” but supplies neither benchmark names nor any numerical results; this should be expanded to a concise results table even in the abstract.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thoughtful comments on our manuscript. We address each of the major comments below, providing clarifications and indicating where revisions will be made to improve the paper.

read point-by-point responses

  1. Referee: [Experiments] Experiments section (and abstract): the headline claim that ACTS “matches full-thinking performance with substantial token savings” rests on the unverified assumption that an RL-tuned controller initialized on synthetic multi-budget trajectories will generalize to real benchmarks without injecting distribution shifts or new errors into the frozen reasoner’s generation. No train/test split of benchmarks, no ablation isolating the RL stage, and no quantitative measure of generation continuity (e.g., token-level divergence or error-injection rate) are supplied, rendering the central empirical result impossible to assess.

    Authors: The benchmarks in our experiments are standard evaluation sets that were not used in constructing the synthetic trajectories, ensuring they serve as held-out test data. The synthetic data is generated from a separate collection of problems with multi-budget augmentation. We agree that an ablation study isolating the RL optimization stage and quantitative measures of generation continuity would provide stronger evidence for the generalization claim. We will include these analyses in the revised manuscript. revision: partial

  2. Referee: [Method] Method section (RL optimization paragraph): the budget-conditioned reward shaping is described only at a high level; without the precise functional form of the reward or the synthetic-data construction procedure, it is impossible to determine whether the learned policy is merely memorizing the augmentation distribution rather than learning transferable steering behavior.

    Authors: While the manuscript describes the approach at a high level, the released code at https://github.com/Andree-9/ACTS contains the full implementation details, including the exact reward function (budget-conditioned combination of accuracy reward and token efficiency penalty) and the procedure for constructing synthetic multi-budget trajectories. To address this, we will add the precise mathematical formulations and a more detailed description of the synthetic data construction to the method section in the revision. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical results from RL training on synthetic data evaluated on external benchmarks

full rationale

The paper formulates steering as an MDP, initializes a controller from constructed synthetic trajectories, optimizes via RL with reward shaping, and reports benchmark results. No step reduces a claimed prediction or result to its own inputs by definition, no fitted parameter is renamed as a prediction, and no self-citation chain bears the central claim. The outcome (token savings with preserved accuracy) is measured on held-out benchmarks rather than being tautological with the training procedure.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The approach rests on the domain assumption that reasoning traces form a Markovian state sufficient for a controller to issue effective steering actions, plus the assumption that synthetic data plus RL will produce a generalizable policy.

axioms (1)
  • domain assumption Reasoning traces plus remaining budget form a Markov state from which a controller can select effective strategy and phrase actions.
    Central to the MDP formulation stated in the abstract.
invented entities (1)
  • Controller agent no independent evidence
    purpose: To observe reasoning state and issue steering actions to a frozen reasoner.
    New component introduced to enable adaptive control.

pith-pipeline@v0.9.1-grok · 5746 in / 1272 out tokens · 31037 ms · 2026-06-28T10:32:12.505848+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

300 extracted references · 141 canonical work pages

  1. [1]

    The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.0

    work page doi:10.18653/v1/2026.wassa-1.0 2026

  2. [2]

    Council of LLM s: Evaluating Capability of Large Language Models to Annotate Propaganda

    Sharma, Vivek and Jain, Shweta and Shokri, Mohammad and Levitan, Sarah Ita and Filatova, Elena. Council of LLM s: Evaluating Capability of Large Language Models to Annotate Propaganda. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.1

    work page doi:10.18653/v1/2026.wassa-1.1 2026

  3. [3]

    Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance

    Tardelli, Serena and Alvisi, Lorenzo and Cima, Lorenzo and Cresci, Stefano and Tesconi, Maurizio. Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.2

    work page doi:10.18653/v1/2026.wassa-1.2 2026

  4. [4]

    Quantifying Social Sentiment in Hostels Using A Domain-Specific Transformer Pipeline

    McMurry, Ian W. Quantifying Social Sentiment in Hostels Using A Domain-Specific Transformer Pipeline. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.3

    work page doi:10.18653/v1/2026.wassa-1.3 2026

  5. [5]

    Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength

    Verma, Bhuvanesh and Marreddy, Mounika and Mehler, Alexander. Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.4

    work page doi:10.18653/v1/2026.wassa-1.4 2026

  6. [6]

    Measuring LLM s' Sensitivity to Paraphrased Opinion Prompts

    Alhetelah, Bushra and Ahmad, Irfan. Measuring LLM s' Sensitivity to Paraphrased Opinion Prompts. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.5

    work page doi:10.18653/v1/2026.wassa-1.5 2026

  7. [7]

    The Impact of Highlighting Subjective Language on Perceived News Trustworthiness

    Shokri, Mohammad and Sharma, Vivek and Klapper, Emily and Jain, Shweta and Filatova, Elena and Levitan, Sarah Ita. The Impact of Highlighting Subjective Language on Perceived News Trustworthiness. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.6

    work page doi:10.18653/v1/2026.wassa-1.6 2026

  8. [8]

    Appraisal Trajectories in Narratives Reveal Distinct Patterns of Emotion Evocation

    Sch. Appraisal Trajectories in Narratives Reveal Distinct Patterns of Emotion Evocation. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.7

    work page doi:10.18653/v1/2026.wassa-1.7 2026

  9. [9]

    Exploring Subjective Tasks in F arsi: A Survey Analysis and Evaluation of Language Models

    Rooein, Donya and Plaza-del-Arco, Flor Miriam and Nozza, Debora and Hovy, Dirk. Exploring Subjective Tasks in F arsi: A Survey Analysis and Evaluation of Language Models. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.8

    work page doi:10.18653/v1/2026.wassa-1.8 2026

  10. [10]

    and Loukachevitch, Natalia V

    Iaroshenko, Polina V. and Loukachevitch, Natalia V. Emotional Lexicons: How Large Language Models Predict Emotional Ratings of R ussian Words. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.9

    work page doi:10.18653/v1/2026.wassa-1.9 2026

  11. [11]

    Emotion-aware text simplification of user generated content using LLM s

    Bezobrazova, Anastasiia and Sokova, Daria and Orasan, Constantin. Emotion-aware text simplification of user generated content using LLM s. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.10

    work page doi:10.18653/v1/2026.wassa-1.10 2026

  12. [12]

    Crowd-Based Evaluation of Emotion Intensity Preservation in S panish -- B asque Tweet Machine Translation

    Aranberri, Nora. Crowd-Based Evaluation of Emotion Intensity Preservation in S panish -- B asque Tweet Machine Translation. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.11

    work page doi:10.18653/v1/2026.wassa-1.11 2026

  13. [13]

    and Markov, Ilia and Vossen, Piek

    Schouten, Stefan F. and Markov, Ilia and Vossen, Piek. A Position Paper on Toxic Reasoning: Grounding Categories of Toxic Language in Implications and Attitudes. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.12

    work page doi:10.18653/v1/2026.wassa-1.12 2026

  14. [14]

    Is Sentiment Banana-Shaped? Exploring the Geometry and Portability of Sentiment Concept Vectors

    Lyngbaek, Laurits and Feldkamp, Pascale and Bizzoni, Yuri and Nielbo, Kristoffer and Enevoldsen, Kenneth. Is Sentiment Banana-Shaped? Exploring the Geometry and Portability of Sentiment Concept Vectors. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/20...

    work page doi:10.18653/v1/2026.wassa-1.13 2026

  15. [15]

    Disentangling Emotion Understanding and Generation in Large Language Models

    Jafari, Sadegh and Lefever, Els and Hoste, Veronique. Disentangling Emotion Understanding and Generation in Large Language Models. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.14

    work page doi:10.18653/v1/2026.wassa-1.14 2026

  16. [16]

    News Credibility Assessment by LLM s and Humans: Implications for Political Bias

    Neves, Pia Wenzel and Jakob, Charlott and Schmitt, Vera. News Credibility Assessment by LLM s and Humans: Implications for Political Bias. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.15

    work page doi:10.18653/v1/2026.wassa-1.15 2026

  17. [17]

    Towards Simulating Social Media Users with LLM s: Evaluating the Operational Validity of Conditioned Comment Prediction

    Schwager, Nils and M. Towards Simulating Social Media Users with LLM s: Evaluating the Operational Validity of Conditioned Comment Prediction. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.16

    work page doi:10.18653/v1/2026.wassa-1.16 2026

  18. [18]

    Label-Consistent Data Generation for Aspect-Based Sentiment Analysis Using LLM Agents

    Monfared, Mohammad Hossein Akbari and Flek, Lucie and Karimi, Akbar. Label-Consistent Data Generation for Aspect-Based Sentiment Analysis Using LLM Agents. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.17

    work page doi:10.18653/v1/2026.wassa-1.17 2026

  19. [19]

    Antisocial Behavior Prediction: A Survey and Practical Guide

    Ollagnier, Ana. Antisocial Behavior Prediction: A Survey and Practical Guide. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.18

    work page doi:10.18653/v1/2026.wassa-1.18 2026

  20. [20]

    Real-Time Mitigation of Negative Emotion in Customer Care Calls

    Gangopadhyay, Surupendu and Mehrabani, Mahnoosh. Real-Time Mitigation of Negative Emotion in Customer Care Calls. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.19

    work page doi:10.18653/v1/2026.wassa-1.19 2026

  21. [21]

    Says Who? Argument Convincingness and Reader Stance Are Correlated with Perceived Author Personality

    Weber, Sabine and Greschner, Lynn and Klinger, Roman. Says Who? Argument Convincingness and Reader Stance Are Correlated with Perceived Author Personality. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.20

    work page doi:10.18653/v1/2026.wassa-1.20 2026

  22. [22]

    A Transformer and Prototype-based Interpretable Model for Contextual Sarcasm Detection

    Wen, Ximing and Rezapour, Rezvaneh. A Transformer and Prototype-based Interpretable Model for Contextual Sarcasm Detection. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.21

    work page doi:10.18653/v1/2026.wassa-1.21 2026

  23. [23]

    Multimodal Claim Extraction for Fact-Checking

    Teo, Joycelyn and Cao, Rui and Deng, Zhenyun and Ding, Zifeng and Schlichtkrull, Michael Sejr and Vlachos, Andreas. Multimodal Claim Extraction for Fact-Checking. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2026.wassa-1.22

    work page doi:10.18653/v1/2026.wassa-1.22 2026

  24. [24]

    A Multi-Aspect Evaluation Framework for Synthetic Data: Case Study on Irony and Sarcasm

    Majer, Laura and Bari \'c , Ana and Sandalj, Florijan and Unkovi \'c , Ivan and Puva c a, Bojan and S najder, Jan. A Multi-Aspect Evaluation Framework for Synthetic Data: Case Study on Irony and Sarcasm. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis ( WASSA 2026). 2026. doi:10.18653/v1/2...

    work page doi:10.18653/v1/2026.wassa-1.23 2026

  25. [25]

    Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.0

    work page doi:10.18653/v1/2026.vardial-1.0 2026

  26. [26]

    and Abdelmoneim, Shahd and Kantharuban, Anjali and Alsboul, Otba and Lamsiyah, Salima and Marchisio, Kelly and Murray, Kenton

    Robinson, Nathaniel R. and Abdelmoneim, Shahd and Kantharuban, Anjali and Alsboul, Otba and Lamsiyah, Salima and Marchisio, Kelly and Murray, Kenton. AMIYA Shared Task: A rabic Modeling In Your Accent at V ar D ial 2026. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.1

    work page doi:10.18653/v1/2026.vardial-1.1 2026

  27. [27]

    Far Out: Evaluating Language Models on Slang in A ustralian and I ndian E nglish

    Dilsiz, Deniz Kaya and Srirag, Dipankar and Joshi, Aditya. Far Out: Evaluating Language Models on Slang in A ustralian and I ndian E nglish. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.2

    work page doi:10.18653/v1/2026.vardial-1.2 2026

  28. [28]

    Effects of Speaker Bias in Dialect Identification and Automatic Transcription with Self-Supervised Speech Models

    Kuparinen, Olli. Effects of Speaker Bias in Dialect Identification and Automatic Transcription with Self-Supervised Speech Models. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.3

    work page doi:10.18653/v1/2026.vardial-1.3 2026

  29. [29]

    O c W iki D ialects: A W ikipedia Dataset With Rich Metadata for O ccitan Dialect Identification

    N \'e dey, Oriane and Bawden, Rachel and Cl \'e rice, Thibault and Sagot, Beno \^i t. O c W iki D ialects: A W ikipedia Dataset With Rich Metadata for O ccitan Dialect Identification. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.4

    work page doi:10.18653/v1/2026.vardial-1.4 2026

  30. [30]

    and Garcia, Marcos

    Irastortza-Urbieta, Xabier and Garc \'i a-Miguel, Jos \'e M. and Garcia, Marcos. Language Mixture to Develop Accurate G alician Dependency Parsers: An Exploration of Its Effects. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.5

    work page doi:10.18653/v1/2026.vardial-1.5 2026

  31. [31]

    Crowdsourcing P iedmontese to Test LLM s on Non-Standard Orthography

    Vico, Gianluca and Libovick \'y , Jind r ich. Crowdsourcing P iedmontese to Test LLM s on Non-Standard Orthography. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.6

    work page doi:10.18653/v1/2026.vardial-1.6 2026

  32. [32]

    G erman- E nglish Code-Switching in Large Language Models

    Aks. G erman- E nglish Code-Switching in Large Language Models. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.7

    work page doi:10.18653/v1/2026.vardial-1.7 2026

  33. [33]

    Perplexity as a Metric for Dialectal Distance: A Computational Study of G reek Varieties

    Chatzikyriakidis, Stergios and Psaltaki, Erofili and Papadakis, Dimitrios and Henriksson, Erik and Laippala, Veronika. Perplexity as a Metric for Dialectal Distance: A Computational Study of G reek Varieties. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.8

    work page doi:10.18653/v1/2026.vardial-1.8 2026

  34. [34]

    A Subword Embedding Approach for Variation Detection in L uxembourgish User Comments

    Lutgen, Anne-Marie and Plum, Alistair and Purschke, Christoph. A Subword Embedding Approach for Variation Detection in L uxembourgish User Comments. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.9

    work page doi:10.18653/v1/2026.vardial-1.9 2026

  35. [35]

    Onomasiological Sense Alignment Across Dialect Dictionaries

    Mederake, Nathalie and Urbach, Nico and Fischer, Hanna and Lameli, Alfred. Onomasiological Sense Alignment Across Dialect Dictionaries. A Taxonomy-Constrained LLM Classification. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.10

    work page doi:10.18653/v1/2026.vardial-1.10 2026

  36. [36]

    and Uban, Ana Sabina and Marchitan, Teodor-George and Iordache, Ioan-Bogdan and Georgescu, Simona

    Dinu, Liviu P. and Uban, Ana Sabina and Marchitan, Teodor-George and Iordache, Ioan-Bogdan and Georgescu, Simona. On the Intelligibility of R omance Language Varieties: S panish and P ortuguese in E urope and A merica. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.11

    work page doi:10.18653/v1/2026.vardial-1.11 2026

  37. [37]

    Dialect Matters: Cross-Lingual ASR Transfer for Low-Resource I ndic Language Varieties

    Dhasmana, Akriti and Srivastava, Aarohi and Chiang, David. Dialect Matters: Cross-Lingual ASR Transfer for Low-Resource I ndic Language Varieties. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.12

    work page doi:10.18653/v1/2026.vardial-1.12 2026

  38. [38]

    Ara- HOPE : Human-Centric Post-Editing Evaluation for Dialectal A rabic to M odern S tandard A rabic Translation

    Alabdullah, Abdullah and Han, Lifeng and Lin, Chenghua. Ara- HOPE : Human-Centric Post-Editing Evaluation for Dialectal A rabic to M odern S tandard A rabic Translation. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.13

    work page doi:10.18653/v1/2026.vardial-1.13 2026

  39. [39]

    I ndic- T uned L ens: Interpreting Multilingual Models in I ndian Languages

    Panchal, Mihir and Varshney, Deeksha and ., Mamta and Ekbal, Asif. I ndic- T uned L ens: Interpreting Multilingual Models in I ndian Languages. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.14

    work page doi:10.18653/v1/2026.vardial-1.14 2026

  40. [40]

    Building ASR Resources for the Hutsul Dialect of U krainian

    Kyslyi, Roman and Orlovskyi, Artem and Khomenko, Pavlo and Onyshchenko, Bohdan and Guzii, Zakhar. Building ASR Resources for the Hutsul Dialect of U krainian. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.15

    work page doi:10.18653/v1/2026.vardial-1.15 2026

  41. [41]

    From F us H a to Folk: Exploring Cross-Lingual Transfer in A rabic Language Models

    Khalak, Abdulmuizz and Issam, Abderrahmane and Spanakis, Gerasimos. From F us H a to Folk: Exploring Cross-Lingual Transfer in A rabic Language Models. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.16

    work page doi:10.18653/v1/2026.vardial-1.16 2026

  42. [42]

    Extending ASR Evaluation Resources for M odern G reek Dialects

    Tsoukala, Chara and Bompolas, Stavros and Margariti, Antigoni and Panagiotou, Konstantina and Plaiti, Maria Elisavet and Tzanakaki, Nefeli and Karatsareas, Petros and Ralli, Angela and Anastasopoulos, Antonios and Markantonatou, Stella. Extending ASR Evaluation Resources for M odern G reek Dialects. Proceedings of the 13th Workshop on NLP for Similar Lang...

    work page doi:10.18653/v1/2026.vardial-1.17 2026

  43. [43]

    How Should We Model the Probability of a Language?

    Dent, Rasul and Ortiz Suarez, Pedro and Cl \'e rice, Thibault and Sagot, Beno \^i t. How Should We Model the Probability of a Language?. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.18

    work page doi:10.18653/v1/2026.vardial-1.18 2026

  44. [44]

    Bridging Dialectal Variation: A Phonetic Transcription Tool for T amil

    Mahaganapathy, Ahrane and Karunakaran, Sumirtha and Navakulan, Kavitha and Sarveswaran, Kengatharaiyer. Bridging Dialectal Variation: A Phonetic Transcription Tool for T amil. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.19

    work page doi:10.18653/v1/2026.vardial-1.19 2026

  45. [45]

    Regional Variation in the Performance of ASR Models on C roatian and S erbian

    Samard z i \'c , Tanja and Rupnik, Peter and Ljube s i \'c , Nikola. Regional Variation in the Performance of ASR Models on C roatian and S erbian. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.20

    work page doi:10.18653/v1/2026.vardial-1.20 2026

  46. [46]

    Syllable Structures Across A rabic Varieties

    Qaddoumi, Abdelrahim and Kodner, Jordan and Khalifa, Salam and Broselow, Ellen and Rambow, Owen. Syllable Structures Across A rabic Varieties. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.21

    work page doi:10.18653/v1/2026.vardial-1.21 2026

  47. [47]

    Curriculum Learning and Pseudo-Labeling Improve the Generalization of Multi-Label A rabic Dialect Identification Models

    Mekky, Ali and El Zeftawy, Mohamed and Hassan, Lara and Keleg, Amr and Nakov, Preslav. Curriculum Learning and Pseudo-Labeling Improve the Generalization of Multi-Label A rabic Dialect Identification Models. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.22

    work page doi:10.18653/v1/2026.vardial-1.22 2026

  48. [48]

    O pen LID -v3: Improving the Precision of Closely Related Language Identification -- An Experience Report

    Fedorova, Mariia and Arefyev, Nikolay and Buljan, Maja and Helcl, Jind r ich and Oepen, Stephan and R nningstad, Egil and Scherrer, Yves. O pen LID -v3: Improving the Precision of Closely Related Language Identification -- An Experience Report. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/202...

    work page doi:10.18653/v1/2026.vardial-1.23 2026

  49. [49]

    Improving Dialect Robustness in Large Language Models via L o RA and Mixture-of-Experts

    Maheshwari, Sanjh and Rajpoot, Aniket Singh and Cocarascu, Oana and ., Mamta. Improving Dialect Robustness in Large Language Models via L o RA and Mixture-of-Experts. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.24

    work page doi:10.18653/v1/2026.vardial-1.24 2026

  50. [50]

    Evaluation Framework for Transfer Learning between Closely Related Lects: A Case Study of Lemko

    Afanasev, Ilia. Evaluation Framework for Transfer Learning between Closely Related Lects: A Case Study of Lemko. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.25

    work page doi:10.18653/v1/2026.vardial-1.25 2026

  51. [51]

    Do Large Language Models Adapt to Language Variation across Socioeconomic Status?

    Bassignana, Elisa and Zhang, Mike and Hovy, Dirk and Cercas Curry, Amanda. Do Large Language Models Adapt to Language Variation across Socioeconomic Status?. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.26

    work page doi:10.18653/v1/2026.vardial-1.26 2026

  52. [52]

    Aladdin- FTI @ AMIYA Three Wishes for A rabic NLP : Fidelity, Diglossia, and Multidialectal Generation

    Mutal, Jonathan and Al Almaoui, Perla and Hengchen, Simon and Bouillon, Pierrette. Aladdin- FTI @ AMIYA Three Wishes for A rabic NLP : Fidelity, Diglossia, and Multidialectal Generation. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.27

    work page doi:10.18653/v1/2026.vardial-1.27 2026

  53. [53]

    Maastricht University at AMIYA : Adapting LLM s for Dialectal A rabic using Fine-tuning and MBR Decoding

    Alali, Abdulhai and Issam, Abderrahmane. Maastricht University at AMIYA : Adapting LLM s for Dialectal A rabic using Fine-tuning and MBR Decoding. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.28

    work page doi:10.18653/v1/2026.vardial-1.28 2026

  54. [54]

    SDNLP at AMIYA 2026: S yrian A rabic Dialect Modeling with L o RA

    Alkhder, Hasan and Abboush, Mohammad. SDNLP at AMIYA 2026: S yrian A rabic Dialect Modeling with L o RA. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.29

    work page doi:10.18653/v1/2026.vardial-1.29 2026

  55. [55]

    NUS - IDS at AMIYA / V ar D ial 2026: Improving A rabic Dialectness in LLM s with Reinforcement Learning

    Gollapalli, Sujatha Das and Hakam, Mouad and Du, Mingzhe and Ng, See-Kiong. NUS - IDS at AMIYA / V ar D ial 2026: Improving A rabic Dialectness in LLM s with Reinforcement Learning. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.30

    work page doi:10.18653/v1/2026.vardial-1.30 2026

  56. [56]

    MBZUAI at AMIYA Shared Task 2026: Adapting Open-Source LLM s for Dialectal A rabic

    Gaber, Rana and Allam, Yara and Amin, Serag and Aly, Ranwa and Alhafni, Bashar. MBZUAI at AMIYA Shared Task 2026: Adapting Open-Source LLM s for Dialectal A rabic. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.31

    work page doi:10.18653/v1/2026.vardial-1.31 2026

  57. [57]

    A Closed-Track System for Palestinian A rabic in the AMIYA Shared Task

    Hamad, Khaleel and Al-Najjar, Ahmad. A Closed-Track System for Palestinian A rabic in the AMIYA Shared Task. Proceedings of the 13th Workshop on NLP for Similar Languages, Varieties and Dialects. 2026. doi:10.18653/v1/2026.vardial-1.32

    work page doi:10.18653/v1/2026.vardial-1.32 2026

  58. [58]

    Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.0

    work page doi:10.18653/v1/2026.teachingnlp-1.0 2026

  59. [59]

    A nimated LLM : Explaining LLM s with Interactive Visualizations

    Kasner, Zden e k and Dusek, Ondrej. A nimated LLM : Explaining LLM s with Interactive Visualizations. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.1

    work page doi:10.18653/v1/2026.teachingnlp-1.1 2026

  60. [60]

    Pedagogic Applications of Argument Maps to Enhance Critical Thinking: Thought Seeds, Argument Mapping, Collaborative Mapping

    Narra, Sruti. Pedagogic Applications of Argument Maps to Enhance Critical Thinking: Thought Seeds, Argument Mapping, Collaborative Mapping. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.2

    work page doi:10.18653/v1/2026.teachingnlp-1.2 2026

  61. [61]

    From Code-Centric to Concept-Centric: Teaching NLP with LLM -Assisted ``Vibe Coding''

    Al-Khalifa, Hend. From Code-Centric to Concept-Centric: Teaching NLP with LLM -Assisted ``Vibe Coding''. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.3

    work page doi:10.18653/v1/2026.teachingnlp-1.3 2026

  62. [62]

    Linguistics to LLM s: Teaching with and about Chatbots

    Pado, Ulrike and Pampel, Barbara. Linguistics to LLM s: Teaching with and about Chatbots. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.4

    work page doi:10.18653/v1/2026.teachingnlp-1.4 2026

  63. [63]

    Language Technology Initiative: Framework for Teaching NLP and Computational Linguistics at the Universities in L atvia

    Skadina, Inguna and Kuzmina, Jana and Platonova, Marina and Smirnova, Tatjana and Kruk, Sergei. Language Technology Initiative: Framework for Teaching NLP and Computational Linguistics at the Universities in L atvia. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.5

    work page doi:10.18653/v1/2026.teachingnlp-1.5 2026

  64. [64]

    Teaching NLP in the AI Era: Experiences from the U niversity of L atvia

    Skadina, Inguna and Barzdins, Guntis and Boj \= a rs, Uldis and Gruzitis, Normunds and Paikens, P \= e teris. Teaching NLP in the AI Era: Experiences from the U niversity of L atvia. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.6

    work page doi:10.18653/v1/2026.teachingnlp-1.6 2026

  65. [65]

    A Hands-on Approach to NLP Fundamentals for External Domain Experts in the LLM Era

    Daza, Angel. A Hands-on Approach to NLP Fundamentals for External Domain Experts in the LLM Era. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.7

    work page doi:10.18653/v1/2026.teachingnlp-1.7 2026

  66. [66]

    and Chervyakov, Artem and Zaytsev, Alexey and Panchenko, Alexander

    Tikhonova, Maria and Chekalina, Viktoriia A. and Chervyakov, Artem and Zaytsev, Alexey and Panchenko, Alexander. From Standard Transformers to M odern LLM s: Bringing Dialogue Models, RAG , and Agents to the Classroom. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.8

    work page doi:10.18653/v1/2026.teachingnlp-1.8 2026

  67. [67]

    Which course? Discourse! Teaching Discourse and Generation in the Era of LLM s

    Li, Junyi Jessy and Liu, Yang Janet and Misra, Kanishka and Pyatkin, Valentina and Sheffield, William. Which course? Discourse! Teaching Discourse and Generation in the Era of LLM s. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.9

    work page doi:10.18653/v1/2026.teachingnlp-1.9 2026

  68. [68]

    From Mixed Backgrounds to NLP Skills

    Barak, Libby and Feldman, Anna. From Mixed Backgrounds to NLP Skills. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.10

    work page doi:10.18653/v1/2026.teachingnlp-1.10 2026

  69. [69]

    Teaching and Critiquing Conceptualization and Operationalization in NLP

    Gautam, Vagrant. Teaching and Critiquing Conceptualization and Operationalization in NLP. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.11

    work page doi:10.18653/v1/2026.teachingnlp-1.11 2026

  70. [70]

    Bridging Applied Experience and Research Contexts in U krainian NLP Education

    Paniv, Yurii and Makovska, Viktoriia. Bridging Applied Experience and Research Contexts in U krainian NLP Education. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.12

    work page doi:10.18653/v1/2026.teachingnlp-1.12 2026

  71. [71]

    Teaching M odern NLP and LLM s at Kyiv School of Economics: A Practice-Oriented Course with U krainian Language Focus

    Kyslyi, Roman and Bazdyrev, Anton. Teaching M odern NLP and LLM s at Kyiv School of Economics: A Practice-Oriented Course with U krainian Language Focus. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.13

    work page doi:10.18653/v1/2026.teachingnlp-1.13 2026

  72. [72]

    Practising responsibility: Ethics in NLP as a hands-on course

    Nissim, Malvina and Patti, Viviana and Savoldi, Beatrice. Practising responsibility: Ethics in NLP as a hands-on course. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.14

    work page doi:10.18653/v1/2026.teachingnlp-1.14 2026

  73. [73]

    Beyond Passive Viewing: A Pilot Study of a Hybrid Learning Platform Augmenting Video Lectures with Conversational AI

    Abraar, Mohammed and Dandekar, Raj and Dandekar, Rajat and Panat, Sreedath. Beyond Passive Viewing: A Pilot Study of a Hybrid Learning Platform Augmenting Video Lectures with Conversational AI. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.15

    work page doi:10.18653/v1/2026.teachingnlp-1.15 2026

  74. [74]

    From Sentiment to Interpretation: Teaching NLP for Literary Understanding Across Educational Contexts

    Bilstrup, Karl-Emil Kj r and Degn, Kirstine Nielsen and Schultz, Morten and Conroy, Alexander and Bjerring-Hansen, Jens and Hershcovich, Daniel. From Sentiment to Interpretation: Teaching NLP for Literary Understanding Across Educational Contexts. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10....

    work page doi:10.18653/v1/2026.teachingnlp-1.16 2026

  75. [75]

    Novel or Drivel? Variants of Invariants for Teaching NLP in the LLM Era

    Micluța-C \^a mpeanu, Marius. Novel or Drivel? Variants of Invariants for Teaching NLP in the LLM Era. Proceedings of the Seventh Workshop on Teaching Natural Language Processing ( T each NLP 2026). 2026. doi:10.18653/v1/2026.teachingnlp-1.17

    work page doi:10.18653/v1/2026.teachingnlp-1.17 2026

  76. [76]

    A ctive LLM : Large Language Model-Based Active Learning for Textual Few-Shot Scenarios

    Bayer, Markus and Lutz, Justin and Reuter, Christian. A ctive LLM : Large Language Model-Based Active Learning for Textual Few-Shot Scenarios. Transactions of the Association for Computational Linguistics. 2026. doi:10.1162/tacl.a.63

    work page doi:10.1162/tacl.a.63 2026

  77. [77]

    M o N a C o: More Natural and Complex Questions for Reasoning Across Dozens of Documents

    Wolfson, Tomer and Trivedi, Harsh and Geva, Mor and Goldberg, Yoav and Roth, Dan and Khot, Tushar and Sabharwal, Ashish and Tsarfaty, Reut. M o N a C o: More Natural and Complex Questions for Reasoning Across Dozens of Documents. Transactions of the Association for Computational Linguistics. 2026. doi:10.1162/tacl.a.64

    work page doi:10.1162/tacl.a.64 2026

  78. [78]

    D eep T rans: Deep Reasoning Translation via Reinforcement Learning

    Wang, Jiaan and Meng, Fandong and Zhou, Jie. D eep T rans: Deep Reasoning Translation via Reinforcement Learning. Transactions of the Association for Computational Linguistics. 2026. doi:10.1162/tacl.a.65

    work page doi:10.1162/tacl.a.65 2026

  79. [79]

    C oref I nst: Leveraging LLM s for Multilingual Coreference Resolution

    Pamay Arslan, Tu. C oref I nst: Leveraging LLM s for Multilingual Coreference Resolution. Transactions of the Association for Computational Linguistics. 2026. doi:10.1162/tacl.a.593

    work page doi:10.1162/tacl.a.593 2026

  80. [80]

    and Josyula, Yasasvi and Choi, Jinho D

    Finch, James D. and Josyula, Yasasvi and Choi, Jinho D. Generative Induction of Dialogue Task Schemas with Streaming Refinement and Simulated Interactions. Transactions of the Association for Computational Linguistics. 2026. doi:10.1162/tacl.a.66

    work page doi:10.1162/tacl.a.66 2026

Showing first 80 references.