Pith. sign in

REVIEW 3 major objections 5 minor 78 references

AMELIA: A Family of Multi-task End-to-end Language Models for Argumentation

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A single Llama-3.1-8B-Instruct model fine-tuned jointly on eight argument-mining tasks built from 19 converted datasets matches or beats every task-specific specialist and sets the paper's top scores on four tasks.

desk verdict Useful unified benchmark and a credible internal comparison of fine-tuning vs merging for argument mining, but the 'state-of-the-art' claim is not supported by the evaluation setup. read the letter →

arxiv 2508.17926 v1 pith:N6KZBDXH submitted 2025-08-25 cs.CL cs.AI

Henri Savigny , Bruno Yun This is my paper · ORCID
classification cs.CLcs.AI
keywords argumentminingmulti-taskfine-tuningmodelmerginglargelanguagemodelsLiaraLlama-3.1-8B-Instructqualityassessmentfallacydetection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Argument mining is normally tackled task by task: one model finds claims, another labels relations, a third scores quality. This paper asks whether one large language model can do all of it, and builds the test bed to find out: it converts 19 existing annotated corpora into one common format, defines eight argument-mining tasks, and fine-tunes Llama-3.1-8B-Instruct on each task, on all tasks at once, and through model merging. Its central claim is that joint training works: the multi-task model matches or beats every specialist (coming within about one and a half points on the tasks where it is not best) and posts the paper's top scores on argument component classification, relation classification, stance detection, and argument quality. If that holds, the practical payoff is that a single model, trained once, can serve most of the argument-mining pipeline, and the merging route makes a cheaper multi-task model available when joint training is too costly.

What carries the argument

Three pieces carry the argument. First, the unified dataset: nineteen corpora with different formats and label inventories are converted into one JSONL schema with a shared label set — for instance Microtext's four relation types ('reb', 'sup', 'und', 'add') are collapsed into support, attack, or no relation — and per-class balanced sampling gives each task 4,000 training, 800 validation, and 800 test instances while preserving each corpus's share. Second, a fixed prompt template (task description, explicit label list, and a <|ANSWER|> delimiter) is used for fine-tuning and inference, so all tasks share one input/output interface, with LoRA (low-rank adaptation, a cheap partial fine-tuning method, rank 16) as the training engine. Third, for the merging route, the paper combines the eight specialists' parameter differences, or task vectors, using the DARE and DELLA methods; DELLA prunes each task vector with probabilities biased toward large-magnitude parameters and then rescales the survivors, the mechanism credited with keeping the hard tasks (fallacies, quality, evidence type) alive in the merged model.

What would settle it

Evaluate the multi-task model on each source dataset's original annotation scheme and native test split instead of the unified conversion — for example, on Microtext's original relation types ('reb', 'sup', 'und', 'add') rather than the collapsed support/attack classes. If the model's scores on the original tasks track the task-specific fine-tuned models, the transfer claim stands; if they drop, the unified benchmark inflated the result.

Watch

Extended reading notes

Core claim

The paper's central claim, stated on its own terms: training one Llama-3.1-8B-Instruct model simultaneously on all eight argument-mining tasks, using 4,000 balanced samples per task drawn from nineteen unified datasets, produces a model that is the best overall performer in its experiments. It reaches 90.74% F1 on argument component classification, 88.33% on argument relation classification, 95.75% on stance detection, and 69.80% on argument quality assessment — the top scores in the paper — and stays within at most about one and a half points of the task-specific fine-tuned models on the other four tasks. The authors read this as evidence that the tasks transfer into each other: structural siblings such as component, claim, evidence-type, and stance tasks share representations, and joint training creates no measurable conflict. They also show that merging the eight specialists with the DELLA method yields a competitive multi-task model (63.64% mean F1) without any joint training, a compromise that trails the jointly trained model mainly on the hardest tasks.

Load-bearing premise

The whole result rests on the assumption that the same label — 'claim', 'premise', 'support', 'attack' — denotes the same thing in all 19 datasets, so that pouring them into one label scheme trains the model on consistent signals rather than on conflicting definitions of the same word.

Editorial extensions

If this is right

  • One jointly fine-tuned Llama-3.1-8B-Instruct can stand in for eight separate argument-mining models: on no task does it fall more than about a point and a half behind the specialist, and on four tasks it is the paper's best.
  • Because joint training shows no degradation, adding further argumentation tasks to the same model is plausible without a dedicated specialist; the measured transfer suggests shared structure among classification-style argument-mining tasks.
  • Merging with difficulty-aware hyperparameters (higher density and weight for hard tasks) gives a multi-task model at a fraction of the training cost, so teams with limited compute or scarce task data have a workable route to a single multi-task system.
  • The quantized GGUF versions keep most of the full-precision performance (94.50% vs 95.75% on stance detection), so the multi-task model can run on modest local hardware.
  • The released unified dataset makes the eight tasks and the conversion rules that produced them reproducible, so other models can be compared on the same balanced splits.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The load-bearing unmeasured factor is label equivalence: the paper collapses each corpus's scheme into one shared vocabulary without measuring how far the original definitions diverge; keeping the full relation inventory distinct (e.g., Microtext's 'reb' vs 'und') would directly test whether the collapsed support/attack labels hide distinctions the benchmark cannot see.
  • The sampling design makes some datasets dominate a task's training mix — IAM Claim alone supplies 1,659 of the 2,000 claim samples in claim detection — so the reported gains may be driven by one or two large corpora; per-dataset reporting would show which corpora actually transfer.
  • A cheap follow-up the paper's modularity claim invites: fine-tune a ninth task's LoRA adapter, merge it into the existing collection, and measure whether the merged model retains the eight old tasks while gaining the ninth; this would test the promised incremental-update property directly.
  • Because task-specific fine-tuning with only 4,000 samples already lifts the model far above its zero-shot baseline on every task, the results hint that strong base models plus small targeted corpora may suffice to adapt an LLM to a new argumentation task — relevant to the many argument-mining domains with little annotated data.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper introduces AMELIA, a family of Llama-3.1-8B-Instruct models for eight argument-mining classification tasks built by converting 19 existing argument-mining datasets into a unified JSONL format. It compares zero-shot and few-shot prompting, task-specific LoRA fine-tuning, joint multi-task fine-tuning, and model merging with DARE and DELLA. The authors report that task-specific fine-tuning improves over baselines, that multi-task fine-tuning preserves or improves per-task performance, and that the DELLA II merge provides a competitive resource-efficient alternative. They further claim state-of-the-art results on argument component classification, argument relation classification, stance detection, and argument quality assessment.

Significance. The paper's main positive contributions are the publicly released converted datasets and model collection, the systematic internal comparison of fine-tuning strategies on a common benchmark, and the exploration of model merging for multi-task argument mining. If the claims are read as internal comparisons on the proposed benchmark, the paper provides useful empirical evidence that multi-task fine-tuning does not degrade performance and that merging is a viable cheap alternative. The paper is open about its resources and describes the merge configurations in enough detail to be reproducible. However, the external state-of-the-art claim is not anchored to shared evaluation splits, and the label-unification step is insufficiently validated; both issues must be addressed before the headline claims can be accepted.

major comments (3)
  1. [Sections 3.3 and 4.3] The 'state-of-the-art' claim is not supported because the evaluation protocol is not shared with the literature. The test sets used here are random 60/20/20 splits of each corpus, further sampled to class-balanced 800-instance subsets, rather than the official test splits or label distributions of the original datasets. Macro-F1 measured on these modified subsets is not comparable to previously published numbers on ACC, AR, SD, or AQ. Table 2 compares only against zero-shot, few-shot, DeBERTa, and the authors' own task-specific models, not against published argument-mining systems on the original tasks. Please either remove or reword the state-of-the-art statements, or evaluate on the original test splits and compare against published numbers.
  2. [Sections 3.1 and 3.2] The unified dataset conversion assumes that labels such as claim, premise, support, attack, and evidence are semantically equivalent across the 19 source corpora, but the annotation guidelines differ substantially. For example, Microtext relation types 'reb', 'sup', 'und', and 'add' are collapsed into the generic support/attack/no-relation scheme without an explicit mapping, and stance labels from FEVER, IBM Claim-polarity, and IAM are treated as one label set despite different annotation conventions. No analysis of label consistency, per-dataset performance, or annotation-schema divergence is provided. This matters because the reported multi-task transfer gains could reflect conflicting or overlapping label conventions rather than shared argumentative structure. Please add per-dataset results and a label-alignment analysis.
  3. [Tables 2 and 4, Section 4.2] All reported numbers come from a single run, so no variance estimates or significance tests support the claims of 'significant' improvement or the differences between merge configurations. In addition, the best merged model (DELLA II) is selected by test-set mean F1 among eight configurations, so the reported 'Merged Model' performance is optimistic relative to a fair model-selection procedure. Please provide multiple seeds with confidence intervals and use a separate validation split for selecting merge hyperparameters before reporting test performance.
minor comments (5)
  1. [Section 3.3] The task set is written as '{ADUC, CD, ED, ARC, ET, SD, FD, AQ}', but the paper elsewhere uses ACC, AR, and FD Single/FD Multi; please make the abbreviations consistent.
  2. [Section 3.2] Example 8 contains the typo 'Ouput', and the fallacy label list includes 'false dillema' instead of 'false dilemma'; also 'ComARg' appears inconsistently as 'ComArg'.
  3. [Section 3.2 and Table 5] The FD Multi evaluation metric is under-specified: the text says a prediction is correct if it belongs to the set of ground-truth fallacies, but it does not define how precision is computed when multiple predictions are produced for one instance. Please state the exact multi-label precision/recall formulas.
  4. [Section 4.1] The DeBERTa baseline is not described in enough detail: it is unclear whether it was fine-tuned on the same training subsets, with which hyperparameters, or on the original tasks. Please specify this to support the comparison.
  5. [Appendix A] Table 5 has formatting artifacts such as 'STM erged' instead of 'ST Merged' and many rows with run-together values; please format the table so that F1, precision, and recall are readable.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the empirical comparisons use held-out splits, and the only self-citations are background references that do not support the load-bearing claims.

full rationale

No load-bearing circular step is present. The paper's central empirical claims are evaluated on held-out test samples (§3.3) taken after a fixed 60/20/20 split of each source corpus, so the reported multi-task and merging results are not defined in terms of the quantities they are used to demonstrate. The label unification in §3.1 is a handcrafted conversion, not a derivation that presupposes the target result. The difficulty split used to configure merging hyperparameters (§4.2) is selected from the same evaluation table, which is a hyperparameter-selection concern rather than a circular prediction: it is an empirical tuning choice, not a quantity that the experiments are claimed to derive from first principles. The self-citations (refs. [4], [6], [8], [10], [18]) appear only in background discussion of formal argumentation and prior LLM-assisted tools, and none is invoked to justify the AMELIA architecture, the dataset conversion, or the reported scores. The 'state-of-the-art' claim in §4.3 is not anchored to the original published test splits because the evaluation uses custom class-balanced subsets, but that is a comparability and correctness issue, not a reduction of the result to its own inputs.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The empirical claims rest on a pipeline with several hand-chosen hyperparameters (LoRA rank, sample sizes, thresholds) and on merging hyperparameters that were tuned against the reported test results. No new theoretical entities are introduced. The main unstated load-bearing assumptions are cross-dataset label equivalence and representativeness of the sampled test sets.

free parameters (5)
  • LoRA rank = 16
    Chosen ad hoc in Section 4.1; no sweep reported, despite the conclusion mentioning future plans to try ranks 32, 64, 128.
  • Training epochs and batch size = 2 epochs, batch size 32
    Fixed for all fine-tuning runs in Section 4.1 without justification or sensitivity analysis.
  • Sample sizes per task = 4000 train, 800 validation, 800 test
    Defined in Section 3.3; these caps, combined with class-balanced sampling, determine the benchmark and all reported scores.
  • Difficulty thresholds for task categorization = 60% and 50%
    Section 4.2 defines easy/medium/hard from the fine-tuned models' own test performance, then assigns merge hyperparameters by difficulty.
  • DELLA II merge hyperparameters = rho 0.9/0.7/0.5, epsilon 0.1/0.15/0.4, w 0.2/0.15/0.03 (hard/medium/easy)
    Table 3; the best configuration is selected by mean F1 on the test set (Section 4.2, Table 4), making these values fitted to the reported evaluation.
assumptions (5)
  • domain assumption Claim, premise, attack, support, and evidence labels are semantically consistent across the 19 datasets after conversion to the unified format.
    Section 3.1 and Example 3 map datasets with different annotation guidelines to a single schema without discussing label divergence.
  • domain assumption The random 60/20/20 split of each dataset and the class-balanced sampling do not leak information between train, validation, and test, and no source text appears in both training and test through different datasets.
    Section 3.3 describes the splits and sampling but does not check overlap between corpora (e.g., IAM and IBM claims may draw from similar web sources).
  • domain assumption A class-balanced 800-sample test subset is representative of each full test split and supports comparisons to published benchmarks.
    Section 3.3 fixes the test size at 800; the paper later uses these scores for 'state-of-the-art' claims without comparing on the original splits.
  • domain assumption The few-shot baseline with one example per label is a fair baseline even for the 20-class fallacy task and the 15-dimension quality task.
    Section 4.1 states the few-shot setup but provides no example selection details or variance across prompts.
  • ad hoc to paper Selecting the reported best merged model (DELLA II) by test-set mean F1 does not materially inflate the reported advantage of merging.
    Section 4.2 and Table 4 select DELLA II after evaluating 8 configurations on the same test sets; no holdout validation for configuration choice is described.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AMELIA: A Family of Multi-task End-to-end Language Models for Argumentation." pith.science (2026). https://pith.science/paper/N6KZBDXH

@misc{pith2026250817926,
  author       = {Pith},
  title        = {Pith review of: AMELIA: A Family of Multi-task End-to-end Language Models for Argumentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N6KZBDXH}},
  note         = {Machine review of arXiv:2508.17926}
}
read the original abstract

Argument mining is a subfield of argumentation that aims to automatically extract argumentative structures and their relations from natural language texts. This paper investigates how a single large language model can be leveraged to perform one or several argument mining tasks. Our contributions are two-fold. First, we construct a multi-task dataset by surveying and converting 19 well-known argument mining datasets from the literature into a unified format. Second, we explore various training strategies using Meta AI's Llama-3.1-8B-Instruct model: (1) fine-tuning on individual tasks, (2) fine-tuning jointly on multiple tasks, and (3) merging models fine-tuned separately on individual tasks. Our experiments show that task-specific fine-tuning significantly improves individual performance across all tasks. Moreover, multi-task fine-tuning maintains strong performance without degradation, suggesting effective transfer learning across related tasks. Finally, we demonstrate that model merging offers a viable compromise: it yields competitive performance while mitigating the computational costs associated with full multi-task fine-tuning.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

78 extracted references · 47 canonical work pages

  1. [1]

    On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic programming and n-person games

    Phan Minh Dung. On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic programming and n-person games. Artificial intelligence, 77(2):321–357, 1995

    1995

  2. [2]

    An introduction to argumentation semantics

    Pietro Baroni, Martin Caminada, and Massimiliano Giacomin. An introduction to argumentation semantics. The knowledge engineering review, 26(4):365–410, 2011

    work page 2011

  3. [3]

    What can argumentation do for inconsistent ontology query answering? In Weiru Liu, V

    Madalina Croitoru and Srdjan Vesic. What can argumentation do for inconsistent ontology query answering? In Weiru Liu, V . S. Subrahmanian, and Jef Wijsen, editors,Scalable Uncertainty Management - 7th International Conference, SUM 2013, Washington, DC, USA, September 16-18, 2013. Proceedings, volume 8078 of Lecture Notes in Computer Science, pages 15–29....

    work page 2013

  4. [4]

    Representing pure nash equilibria in argumentation

    Bruno Yun, Srdjan Vesic, and Nir Oren. Representing pure nash equilibria in argumentation. Argument Comput., 13(2):195–208, 2022

    work page 2022

  5. [5]

    On the bipolarity in argumentation frameworks

    Leila Amgoud, Claudette Cayrol, and Marie-Christine Lagasquie-Schiex. On the bipolarity in argumentation frameworks. In James P. Delgrande and Torsten Schaub, editors, 10th International Workshop on Non-Monotonic Reasoning (NMR 2004), Whistler, Canada, June 6-8, 2004, Proceedings, pages 1–9, 2004

    work page 2004

  6. [6]

    Gradual semantics for weighted bipolar setafs

    Bruno Yun and Srdjan Vesic. Gradual semantics for weighted bipolar setafs. In Jirina Vejnarová and Nic Wilson, editors, Symbolic and Quantitative Approaches to Reasoning with Uncertainty - 16th European Conference, ECSQARU 2021, Prague, Czech Republic, September 21-24, 2021, Proceedings, volume 12897 of Lecture Notes in Computer Science, pages 201–214. Sp...

    work page 2021

  7. [7]

    Computing preferred extensions for argumentation systems with sets of attacking arguments

    Søren Holbech Nielsen and Simon Parsons. Computing preferred extensions for argumentation systems with sets of attacking arguments. In Paul E. Dunne and Trevor J. M. Bench-Capon, editors, Computational Models of Argument: Proceedings of COMMA 2006, September 11-12, 2006, Liverpool, UK, volume 144 of Frontiers in Artificial Intelligence and Applications, p...

    work page 2006

  8. [8]

    Sets of attacking arguments for inconsistent datalog knowledge bases

    Bruno Yun, Srdjan Vesic, and Madalina Croitoru. Sets of attacking arguments for inconsistent datalog knowledge bases. In Henry Prakken, Stefano Bistarelli, Francesco Santini, and Carlo Taticchi, editors, Computational Models of Argument - Proceedings of COMMA 2020, Perugia, Italy, September 4-11, 2020, volume 326 of Frontiers in Artificial Intelligence an...

    work page 2020

Show all 78 references
  1. [9]

    Evaluation of argument strength in attack graphs: Foundations and semantics

    Leila Amgoud, Dragan Doder, and Srdjan Vesic. Evaluation of argument strength in attack graphs: Foundations and semantics. Artificial Intelligence, 302:103607, 2022

    2022

  2. [10]

    Ranking-based semantics for sets of attacking arguments

    Bruno Yun, Srdjan Vesic, and Madalina Croitoru. Ranking-based semantics for sets of attacking arguments. InThe Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, The T...

    2020

  3. [11]

    Using arguments for making and explaining decisions

    Leila Amgoud and Henri Prade. Using arguments for making and explaining decisions. Artificial Intelligence, 173(3-4):413–436, 2009

    2009

  4. [12]

    Argument mining: A survey

    John Lawrence and Chris Reed. Argument mining: A survey. Computational Linguistics, 45(4):765–818, 2020

    2020

  5. [13]

    When automated fact-checking meets argumentation: Unveiling fake news through argumentative evidence

    Xiaoou Wang, Elena Cabrio, and Serena Villata. When automated fact-checking meets argumentation: Unveiling fake news through argumentative evidence. Argument & Computation, page 19462174251330980, 2025

    2025

  6. [14]

    Using argumentation theory to fight misinformation

    Blanca Calvo Figueras. Using argumentation theory to fight misinformation. 2024

    2024

  7. [15]

    Argument mining with structured svms and rnns

    Vlad Niculae, Joonsuk Park, and Claire Cardie. Argument mining with structured svms and rnns. arXiv preprint arXiv:1704.06869, 2017

    2017 arXiv

  8. [16]

    Transformer-based argument mining for healthcare applications

    Tobias Mayer, Elena Cabrio, and Serena Villata. Transformer-based argument mining for healthcare applications. In ECAI 2020, pages 2108–2115. IOS Press, 2020. 17 Savigny and Yun

    2020

  9. [17]

    Can large language models perform relation-based argument mining? arXiv preprint arXiv:2402.11243, 2024

    Deniz Gorur, Antonio Rago, and Francesca Toni. Can large language models perform relation-based argument mining? arXiv preprint arXiv:2402.11243, 2024

    2024 arXiv

  10. [18]

    Assisted debate builder with large language models

    Elliot Faugier, Frédéric Armetta, Angela Bonifati, and Bruno Yun. Assisted debate builder with large language models. In ECAI 2024, pages 4447–4450. IOS Press, 2024

    2024

  11. [19]

    Argument mining with fine-tuned large language models

    Jérémie Cabessa, Hugo Hernault, and Umer Mushtaq. Argument mining with fine-tuned large language models. In Proceedings of the 31st International Conference on Computational Linguistics, pages 6624–6635, 2025

    2025

  12. [20]

    Arginstruct: Specialized instruction fine-tuning for computational argumentation

    Maja Stahl, Timon Ziegenbein, Joonsuk Park, and Henning Wachsmuth. Arginstruct: Specialized instruction fine-tuning for computational argumentation. arXiv preprint arXiv:2505.22076, 2025

    2025 arXiv

  13. [21]

    Efficient estimation of word representations in vector space

    Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781, 2013

    2013 arXiv

  14. [22]

    Language models are unsupervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9, 2019

    2019

  15. [23]

    Language models are few-shot learners

    Ben Mann, N Ryder, M Subbiah, J Kaplan, P Dhariwal, A Neelakantan, P Shyam, G Sastry, A Askell, S Agarwal, et al. Language models are few-shot learners. arXiv preprint arXiv:2005.14165, 1:3, 2020

    2005 arXiv

  16. [24]

    The llama 3 herd of models

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783, 2024

    2024 arXiv

  17. [25]

    Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas ...

    2023

  18. [26]

    Qwen3 technical report

    An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, et al. Qwen3 technical report. arXiv preprint arXiv:2505.09388, 2025

    2025 arXiv

  19. [27]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

    2023 arXiv

  20. [28]

    Claude’s extended thinking

    Anthropic. Claude’s extended thinking. https://www.anthropic.com/research/ visible-extended-thinking, 2025

    2025

  21. [29]

    Introducing cogito preview, 2025

    DeepCogito. Introducing cogito preview, 2025

    2025

  22. [30]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025

    2025 arXiv

  23. [31]

    Deepseek-v3 technical report

    Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, et al. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437, 2024

    2024 arXiv

  24. [32]

    QwQ: A High-Performance Reasoning Model for Theorem Proving and Code Generation

    Chengrun Yang, Xuezhi Wang, and et al. QwQ: A High-Performance Reasoning Model for Theorem Proving and Code Generation. arXiv preprint arXiv:2404.07158, 2024

    2024 arXiv

  25. [33]

    Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022

    2022

  26. [34]

    Beyond chain-of-thought, effective graph-of-thought reasoning in language models

    Yao Yao, Zuchao Li, and Hai Zhao. Beyond chain-of-thought, effective graph-of-thought reasoning in language models. arXiv preprint arXiv:2305.16582, 2023

    2023 arXiv

  27. [35]

    Thread of thought unraveling chaotic contexts

    Yucheng Zhou, Xiubo Geng, Tao Shen, Chongyang Tao, Guodong Long, Jian-Guang Lou, and Jianbing Shen. Thread of thought unraveling chaotic contexts. arXiv preprint arXiv:2311.08734, 2023

    2023 arXiv

  28. [36]

    Scaling instruction-finetuned language models

    Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, et al. Scaling instruction-finetuned language models. Journal of Machine Learning Research, 25(70):1–53, 2024

    2024

  29. [37]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3, 2022

    2022

  30. [38]

    Learning from revisions: Quality assessment of claims in argumentation at scale

    Gabriella Skitalinskaya, Jonas Klaff, and Henning Wachsmuth. Learning from revisions: Quality assessment of claims in argumentation at scale. arXiv preprint arXiv:2101.10250, 2021. 18 Savigny and Yun

    2021 arXiv

  31. [39]

    Contextual interaction for argument post quality assessment

    Yiran Wang, Xuanang Chen, Ben He, and Le Sun. Contextual interaction for argument post quality assessment. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 10420–10432, 2023

    2023

  32. [40]

    Computational argumentation quality assessment in natural language

    Henning Wachsmuth, Nona Naderi, Yufang Hou, Yonatan Bilu, Vinodkumar Prabhakaran, Tim Alberdingk Thijm, Graeme Hirst, and Benno Stein. Computational argumentation quality assessment in natural language. In Proceedings of the 15th Conference of the European Chapter of the Assoc...

    2017

  33. [41]

    Towards relation based argumentation mining

    Lucas Carstens and Francesca Toni. Towards relation based argumentation mining. In Proceedings of the 2nd Workshop on Argumentation Mining, pages 29–34, 2015

    2015

  34. [42]

    Classifying argumentative relations using logical mechanisms and argumentation schemes

    Yohan Jo, Seojin Bang, Chris Reed, and Eduard Hovy. Classifying argumentative relations using logical mechanisms and argumentation schemes. Transactions of the Association for Computational Linguistics, 9:721– 739, 2021

    2021

  35. [43]

    Never retreat, never retract: Argumentation analysis for political speeches

    Stefano Menini, Elena Cabrio, Sara Tonelli, and Serena Villata. Never retreat, never retract: Argumentation analysis for political speeches. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32, 2018

    2018

  36. [44]

    Mafalda: A benchmark and comprehensive study of fallacy detection and classification

    Chadi Helwe, Tom Calamai, Pierre-Henri Paris, Chloé Clavel, and Fabian Suchanek. Mafalda: A benchmark and comprehensive study of fallacy detection and classification. arXiv preprint arXiv:2311.09761, 2023

    2023 arXiv

  37. [45]

    Argument-based detection and classifica- tion of fallacies in political debates

    Pierpaolo Goffredo, Mariana Chaves, Serena Villata, and Elena Cabrio. Argument-based detection and classifica- tion of fallacies in political debates. In EMNLP 2023-Conference on Empirical Methods in Natural Language Processing, volume 2023, pages 11101–11112. Association for ...

    2023

  38. [46]

    Towards an argument mining pipeline transforming texts to argument graphs

    Mirko Lenz, Premtim Sahitaj, Sean Kallenberg, Christopher Coors, Lorik Dumani, Ralf Schenkel, and Ralph Bergmann. Towards an argument mining pipeline transforming texts to argument graphs. In Computational Models of Argument, pages 263–270. IOS Press, 2020

    2020

  39. [47]

    End-to-end argument mining with cross- corpora multi-task learning

    Gaku Morio, Hiroaki Ozaki, Terufumi Morishita, and Kohsuke Yanai. End-to-end argument mining with cross- corpora multi-task learning. Transactions of the Association for Computational Linguistics, 10:639–658, 2022

    2022

  40. [48]

    Multi-task learning for argumentation mining in low-resource settings

    Claudia Schulz, Steffen Eger, Johannes Daxenberger, Tobias Kahse, and Iryna Gurevych. Multi-task learning for argumentation mining in low-resource settings. arXiv preprint arXiv:1804.04083, 2018

    2018 arXiv

  41. [49]

    Dataset independent baselines for relation prediction in argument mining

    Oana Cocarascu, Elena Cabrio, Serena Villata, and Francesca Toni. Dataset independent baselines for relation prediction in argument mining. In Computational Models of Argument, pages 45–52. IOS Press, 2020

    2020

  42. [50]

    Argument mining as a text-to- text generation task

    Masayuki Kawarada, Tsutomu Hirao, Wataru Uchida, and Masaaki Nagata. Argument mining as a text-to- text generation task. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2002–2014, 2024

    2002

  43. [51]

    Aqe: argument quadruplet extraction via a quad-tagging augmented generative approach

    Jia Guo, Liying Cheng, Wenxuan Zhang, Stanley Kok, Xin Li, and Lidong Bing. Aqe: argument quadruplet extraction via a quad-tagging augmented generative approach. arXiv preprint arXiv:2305.19902, 2023

    2023 arXiv

  44. [52]

    Assessing open-source large language models on argumentation mining subtasks

    Mohammad Yeghaneh Abkenar, Weixing Wang, Hendrik Graupner, and Manfred Stede. Assessing open-source large language models on argumentation mining subtasks. arXiv preprint arXiv:2411.05639, 2024

    2024 arXiv

  45. [53]

    Are large language models reliable argument quality annotators? In Conference on Advances in Robust Argumentation Machines, pages 129–146

    Nailia Mirzakhmedova, Marcel Gohsen, Chia Hao Chang, and Benno Stein. Are large language models reliable argument quality annotators? In Conference on Advances in Robust Argumentation Machines, pages 129–146. Springer, 2024

    2024

  46. [54]

    Detecting argumentative fallacies in the wild: Problems and limitations of large language models

    Ramon Ruiz-Dolz and John Lawrence. Detecting argumentative fallacies in the wild: Problems and limitations of large language models. In Proceedings of the 10th Workshop on Argument Mining. Association for Computational Linguistics, 2023

    2023

  47. [55]

    Cocolofa: A dataset of news comments with common logical fallacies written by llm-assisted crowds

    Min-Hsuan Yeh, Ruyuan Wan, and Ting-Hao’Kenneth’ Huang. Cocolofa: A dataset of news comments with common logical fallacies written by llm-assisted crowds. arXiv preprint arXiv:2410.03457, 2024

    2024 arXiv

  48. [56]

    In-context learning and fine-tuning gpt for argument mining

    Jérémie Cabessa, Hugo Hernault, and Umer Mushtaq. In-context learning and fine-tuning gpt for argument mining. arXiv preprint arXiv:2406.06699, 2024

    2024 arXiv

  49. [57]

    Which side are you on? a multi-task dataset for end-to-end argument summarisation and evaluation

    Hao Li, Yuping Wu, Viktor Schlegel, Riza Batista-Navarro, Tharindu Madusanka, Iqra Zahid, Jiayan Zeng, Xiaochi Wang, Xinran He, Yizhi Li, et al. Which side are you on? a multi-task dataset for end-to-end argument summarisation and evaluation. arXiv preprint arXiv:2406.03151, 2024

    2024 arXiv

  50. [58]

    Back up your stance: Recognizing arguments in online discussions

    Filip Boltuži´c and Jan Šnajder. Back up your stance: Recognizing arguments in online discussions. In Proceedings of the first workshop on argumentation mining, pages 49–58, 2014

    2014

  51. [59]

    Fever: a large-scale dataset for fact extraction and verification

    James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. Fever: a large-scale dataset for fact extraction and verification. arXiv preprint arXiv:1803.05355, 2018. 19 Savigny and Yun

    2018 arXiv

  52. [60]

    Iam: a comprehensive and large-scale dataset for integrated argument mining tasks

    Liying Cheng, Lidong Bing, Ruidan He, Qian Yu, Yan Zhang, and Luo Si. Iam: a comprehensive and large-scale dataset for integrated argument mining tasks. arXiv preprint arXiv:2203.12257, 2022

    2022 arXiv

  53. [61]

    Stance classification of context-dependent claims

    Roy Bar-Haim, Indrajit Bhattacharya, Francesco Dinuzzo, Amrita Saha, and Noam Slonim. Stance classification of context-dependent claims. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages...

    2017

  54. [62]

    A benchmark dataset for automatic detection of claims and evidence in the context of controversial topics

    Ehud Aharoni, Anatoly Polnarov, Tamar Lavee, Daniel Hershcovich, Ran Levy, Ruty Rinott, Dan Gutfreund, and Noam Slonim. A benchmark dataset for automatic detection of claims and evidence in the context of controversial topics. In Proceedings of the first workshop on argumentat...

    2014

  55. [63]

    Towards an argumentative content search engine using weak supervision

    Ran Levy, Ben Bogin, Shai Gretz, Ranit Aharonov, and Noam Slonim. Towards an argumentative content search engine using weak supervision. In Proceedings of the 27th International Conference on Computational Linguistics, pages 2066–2081, 2018

    2018

  56. [64]

    Will it blend? blending weak and strong labeled data in a neural network for argumentation mining

    Eyal Shnarch, Carlos Alzate, Lena Dankin, Martin Gleize, Yufang Hou, Leshem Choshen, Ranit Aharonov, and Noam Slonim. Will it blend? blending weak and strong labeled data in a neural network for argumentation mining. In Proceedings of the 56th Annual Meeting of the Association...

    2018

  57. [65]

    Unsupervised expressive rules provide explainability and assist human experts grasping new domains

    Eyal Shnarch, Leshem Choshen, Guy Moshkowich, Noam Slonim, and Ranit Aharonov. Unsupervised expressive rules provide explainability and assist human experts grasping new domains. arXiv preprint arXiv:2010.09459, 2020

    2010 arXiv

  58. [66]

    An annotated corpus of argumentative microtexts

    Andreas Peldszus and Manfred Stede. An annotated corpus of argumentative microtexts. In Argumentation and Reasoned Action: Proceedings of the 1st European Conference on Argumentation, Lisbon, volume 2, pages 801–815, 2015

    2015

  59. [67]

    More or less controlled elicitation of argumentative text: Enlarging a microtext corpus via crowdsourcing

    Maria Skeppstedt, Andreas Peldszus, and Manfred Stede. More or less controlled elicitation of argumentative text: Enlarging a microtext corpus via crowdsourcing. In Proceedings of the 5th Workshop on Argument Mining, pages 155–163, 2018

    2018

  60. [68]

    Node: A benchmark of natural language arguments

    Elena Cabrio and Serena Villata. Node: A benchmark of natural language arguments. In Computational Models of Argument, pages 449–450. IOS Press, 2014

    2014

  61. [69]

    Parsing argumentation structures in persuasive essays

    Christian Stab and Iryna Gurevych. Parsing argumentation structures in persuasive essays. Computational Linguistics, 43(3):619–659, 2017

    2017

  62. [70]

    Deberta: Decoding-enhanced bert with disentan- gled attention

    Pengcheng He, Xiaodong Liu, Jianfeng Gao, and Weizhu Chen. Deberta: Decoding-enhanced bert with disentan- gled attention. arXiv preprint arXiv:2006.03654, 2020

    2006 arXiv

  63. [71]

    Arcee’s MergeKit: A toolkit for merging large language models

    Charles Goddard, Shamane Siriwardhana, Malikeh Ehghaghi, Luke Meyers, Vladimir Karpukhin, Brian Benedict, Mark McQuade, and Jacob Solawetz. Arcee’s MergeKit: A toolkit for merging large language models. In Franck Dernoncourt, Daniel Preo¸ tiuc-Pietro, and Anastasia Shimorina, ...

    2024

  64. [72]

    Editing models with task arithmetic

    Gabriel Ilharco, Marco Tulio Ribeiro, Mitchell Wortsman, Suchin Gururangan, Ludwig Schmidt, Hannaneh Hajishirzi, and Ali Farhadi. Editing models with task arithmetic. arXiv preprint arXiv:2212.04089, 2022

    2022 arXiv

  65. [73]

    Ties-merging: Resolving interference when merging models

    Prateek Yadav, Derek Tam, Leshem Choshen, Colin A Raffel, and Mohit Bansal. Ties-merging: Resolving interference when merging models. Advances in Neural Information Processing Systems, 36:7093–7115, 2023

    2023

  66. [74]

    Language models are super mario: Absorbing abilities from homologous models as a free lunch

    Le Yu, Bowen Yu, Haiyang Yu, Fei Huang, and Yongbin Li. Language models are super mario: Absorbing abilities from homologous models as a free lunch. In Forty-first International Conference on Machine Learning, 2024

    2024

  67. [75]

    Della-merging: Reducing interference in model merging through magnitude-based sampling

    Pala Tej Deep, Rishabh Bhardwaj, and Soujanya Poria. Della-merging: Reducing interference in model merging through magnitude-based sampling. arXiv preprint arXiv:2406.11617, 2024

    2024 arXiv

  68. [76]

    Rouge: A package for automatic evaluation of summaries

    Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81, 2004

    2004

  69. [77]

    Meteor 1.3: Automatic metric for reliable optimization and evaluation of machine translation systems

    Michael Denkowski and Alon Lavie. Meteor 1.3: Automatic metric for reliable optimization and evaluation of machine translation systems. In Proceedings of the sixth workshop on statistical machine translation, pages 85–91, 2011

    2011

  70. [78]

    Bertscore: Evaluating text generation with bert

    Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675, 2019. 20 Savigny and Yun !hbpModel ACC CD ED AR ET SD FDSingle FDM ulti AQF1 P R F1 P R F1 P R F1 P R F1 P R F1 P R F1...

    1904 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.