ArgBench unifies 33 existing datasets into a standardized benchmark for testing LLMs across 46 argumentation tasks and analyzes the impact of prompting techniques and model factors on performance.
Title resolution pending
48 Pith papers cite this work. Polarity classification is still indexing.
48
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2representative citing papers
ORPO performs preference alignment during supervised fine-tuning via a monolithic odds ratio penalty, allowing 7B models to outperform larger state-of-the-art models on alignment benchmarks.
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
Presents first online L2D algorithm for multiclass classification with bandit feedback and varying experts, achieving O((n+n_e)T^{2/3}) regret generally and O((n+n_e)√T) under low noise.
Introduces TBPO, which derives a Bregman-divergence density-ratio matching objective for token-level preference optimization that generalizes DPO while preserving the induced optimal policy.
OLIVIA treats LLM agent action selection as a contextual linear bandit over frozen hidden states and applies UCB exploration to adapt online, yielding consistent gains over static ReAct and prompt-based baselines on four benchmarks.
LOVER creates an unsupervised logic-regularized verifier that reaches 95% of supervised verifier performance on reasoning tasks across 10 datasets.
LLMs process negation using both attention-based suppression and constructive representation mechanisms (construction dominant), with late-layer attention shortcuts explaining poor accuracy on negation tasks.
S²R² improves robustness of LoRA-tuned LLMs to prompt perturbations by penalizing semantic-segment drift while preserving clean performance and cross-dataset transfer.
Adaptive Instruction Composition uses a neural contextual bandit with RL to adaptively combine crowdsourced texts, generating more effective and diverse LLM jailbreaks than random or prior adaptive methods on Harmbench.
Single-agent systems with tools provide the optimal performance-efficiency trade-off for small language models, outperforming base models and multi-agent setups.
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
M3-Embedding is a single model for multi-lingual, multi-functional, and multi-granular text embeddings trained via self-knowledge distillation that achieves new state-of-the-art results on multilingual, cross-lingual, and long-document retrieval benchmarks.
MAFIG is a multi-agent framework that uses LLM agents and evaluators to generate reading comprehension items with significantly higher adherence to specified feature constraints than single-agent baselines.
Nexa learns a response-conditioned policy that starts with parallel agent execution and adds at most one round of sequential message passing via a predicted sparse DAG, strictly subsuming pure parallel mode.
Dynamically adjusting beta via LLM-as-judge downweights biased comparisons to learn more rational reward models from flawed human preferences.
LaaB improves LLM hallucination detection by mapping self-judgment labels back into neural feature space and using mutual learning under logical consistency constraints between responses and meta-judgments.
MultiBreak is a large diverse multi-turn jailbreak benchmark that achieves substantially higher attack success rates on LLMs than prior datasets and reveals topic-specific vulnerabilities in multi-turn settings.
CleanBase identifies malicious documents in RAG databases by detecting cliques in a semantic similarity graph constructed using embedding models and a statistical threshold.
BREW uses block voting and window-shifting verification to reach TPR 0.965 and FPR 0.02 under 10% synonym substitution, addressing high false-positive issues in prior multi-bit LLM watermarking.
SAMoRA is a parameter-efficient fine-tuning framework that uses semantic-aware routing and task-adaptive scaling within a Mixture of LoRA Experts to improve multi-task performance and generalization over prior methods.
LogosKG delivers a novel hardware-aligned system for efficient multi-hop retrieval on billion-edge knowledge graphs without sacrificing fidelity, demonstrated via biomedical KG-LLM applications.
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.
LLMs prompted with few-shot examples and rationales generate better reasoned distractors for MCQs than fine-tuned contrastive models across six benchmarks.
citing papers explorer
-
ArgBench: Benchmarking LLMs on Computational Argumentation Tasks
ArgBench unifies 33 existing datasets into a standardized benchmark for testing LLMs across 46 argumentation tasks and analyzes the impact of prompting techniques and model factors on performance.
-
ORPO: Monolithic Preference Optimization without Reference Model
ORPO performs preference alignment during supervised fine-tuning via a monolithic odds ratio penalty, allowing 7B models to outperform larger state-of-the-art models on alignment benchmarks.
-
Semantic Reranking at Inference Time for Hard Examples in Rhetorical Role Labeling
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
-
Online Learning-to-Defer with Varying Experts
Presents first online L2D algorithm for multiclass classification with bandit feedback and varying experts, achieving O((n+n_e)T^{2/3}) regret generally and O((n+n_e)√T) under low noise.
-
TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching
Introduces TBPO, which derives a Bregman-divergence density-ratio matching objective for token-level preference optimization that generalizes DPO while preserving the induced optimal policy.
-
OLIVIA: Online Learning via Inference-time Action Adaptation for Decision Making in LLM ReAct Agents
OLIVIA treats LLM agent action selection as a contextual linear bandit over frozen hidden states and applies UCB exploration to adapt online, yielding consistent gains over static ReAct and prompt-based baselines on four benchmarks.
-
Logic-Regularized Verifier Elicits Reasoning from LLMs
LOVER creates an unsupervised logic-regularized verifier that reaches 95% of supervised verifier performance on reasoning tasks across 10 datasets.
-
How Language Models Process Negation
LLMs process negation using both attention-based suppression and constructive representation mechanisms (construction dominant), with late-layer attention shortcuts explaining poor accuracy on negation tasks.
-
Where Do Prompt Perturbations Break Generation? A Segment-Level View of Robustness in LoRA-Tuned Language Models
S²R² improves robustness of LoRA-tuned LLMs to prompt perturbations by penalizing semantic-segment drift while preserving clean performance and cross-dataset transfer.
-
Adaptive Instruction Composition for Automated LLM Red-Teaming
Adaptive Instruction Composition uses a neural contextual bandit with RL to adaptively combine crowdsourced texts, generating more effective and diverse LLM jailbreaks than random or prior adaptive methods on Harmbench.
-
Rethinking Scale: Deployment Trade-offs of Small Language Models under Agent Paradigms
Single-agent systems with tools provide the optimal performance-efficiency trade-off for small language models, outperforming base models and multi-agent setups.
-
SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
-
M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
M3-Embedding is a single model for multi-lingual, multi-functional, and multi-granular text embeddings trained via self-knowledge distillation that achieves new state-of-the-art results on multilingual, cross-lingual, and long-document retrieval benchmarks.
-
A Multi-Agent Framework for Feature-Constrained Difficulty Control in Reading Comprehension Item Generation
MAFIG is a multi-agent framework that uses LLM agents and evaluators to generate reading comprehension items with significantly higher adherence to specified feature constraints than single-agent baselines.
-
Response-Conditioned Parallel-to-Sequential Orchestration for Multi-Agent Systems
Nexa learns a response-conditioned policy that starts with parallel agent execution and adds at most one round of sequential message passing via a predicted sparse DAG, strictly subsuming pure parallel mode.
-
Mitigating Cognitive Bias in RLHF by Altering Rationality
Dynamically adjusting beta via LLM-as-judge downweights biased comparisons to learn more rational reward models from flawed human preferences.
-
Logical Consistency as a Bridge: Improving LLM Hallucination Detection via Label Constraint Modeling between Responses and Self-Judgments
LaaB improves LLM hallucination detection by mapping self-judgment labels back into neural feature space and using mutual learning under logical consistency constraints between responses and meta-judgments.
-
MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety
MultiBreak is a large diverse multi-turn jailbreak benchmark that achieves substantially higher attack success rates on LLMs than prior datasets and reveals topic-specific vulnerabilities in multi-turn settings.
-
CleanBase: Detecting Malicious Documents in RAG Knowledge Databases
CleanBase identifies malicious documents in RAG databases by detecting cliques in a semantic similarity graph constructed using embedding models and a statistical threshold.
-
Block-wise Codeword Embedding for Reliable Multi-bit Text Watermarking
BREW uses block voting and window-shifting verification to reach TPR 0.965 and FPR 0.02 under 10% synonym substitution, addressing high false-positive issues in prior multi-bit LLM watermarking.
-
SAMoRA: Semantic-Aware Mixture of LoRA Experts for Task-Adaptive Learning
SAMoRA is a parameter-efficient fine-tuning framework that uses semantic-aware routing and task-adaptive scaling within a Mixture of LoRA Experts to improve multi-task performance and generalization over prior methods.
-
LogosKG: Hardware-Optimized Scalable and Interpretable Knowledge Graph Retrieval
LogosKG delivers a novel hardware-aligned system for efficient multi-hop retrieval on billion-edge knowledge graphs without sacrificing fidelity, demonstrated via biomedical KG-LLM applications.
-
Characterizing Model-Native Skills
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.
-
Beyond Fine-Tuning: In-Context Learning and Chain-of-Thought for Reasoned Distractor Generation
LLMs prompted with few-shot examples and rationales generate better reasoned distractors for MCQs than fine-tuned contrastive models across six benchmarks.
-
Learning to Control Summaries with Score Ranking
A score-ranking loss enables controllable summarization by aligning outputs to evaluation scores, matching SOTA performance with dimension-specific control on LLaMA, Qwen, and Mistral.
-
Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive
DPOP is a new loss function that prevents DPO from lowering preferred response likelihoods and outperforms standard DPO on diverse datasets, MT-Bench, and enables Smaug-72B to exceed 80% on the Open LLM Leaderboard.
-
Nomic Embed: Training a Reproducible Long Context Text Embedder
Nomic AI produced and open-sourced a reproducible 8192-context English text embedder that exceeds OpenAI Ada-002 and text-embedding-3-small performance on MTEB short-context and LoCo long-context benchmarks.
-
Investigating Cross-Modal Skill Injection: Scenarios, Methods, and Hyperparameters
Systematic evaluation finds cross-modal skill injection via model merging succeeds in instruction-following and cross-lingual scenarios but fails in mathematical reasoning, with TA and DARE methods outperforming others after hyperparameter analysis.
-
Grounded Satirical Generation with RAG
RAG and topic-based word selection increase perceived political relevance in generated satirical definitions but produce no clear improvement in humor according to human raters.
-
PRISM: Perception Reasoning Interleaved for Sequential Decision Making
PRISM interleaves VLM perception and LLM reasoning via a dynamic goal-oriented question-answer pipeline to produce sharper scene descriptions, outperforming prior image-based models on ALFWorld and Room-to-Room.
-
Why Someone Asked "Why": Foil Inference in Human and LLM Question Interpretation
People infer the foil of a why-question mainly from hindsight expectations about what the asker finds surprising after the outcome.
-
Mesh Based Simulations with Spatial and Temporal awareness
A unified training framework for mesh-based ML surrogates in CFD improves accuracy and long-horizon stability by enforcing spatial derivative consistency via multi-node prediction, using temporal cross-attention correction, and adding 3D rotary positional embeddings.
-
From Flat Facts to Sharp Hallucinations: Detecting Stubborn Errors via Gradient Sensitivity
EPGS detects high-confidence factual errors in LLMs by using embedding perturbations to measure gradient sensitivity as a proxy for sharp versus flat minima.
-
GiVA: Gradient-Informed Bases for Vector-Based Adaptation
GiVA uses gradients to initialize vector adapters so they match LoRA performance at eight times lower rank while keeping extreme parameter efficiency.
-
GS-Quant: Granular Semantic and Generative Structural Quantization for Knowledge Graph Completion
GS-Quant generates coarse-to-fine discrete codes for KG entities via semantic hierarchy injection and causal sequence reconstruction, enabling LLMs to perform knowledge graph completion by treating the codes as vocabulary tokens.
-
Can Continual Pre-training Bridge the Performance Gap between General-purpose and Specialized Language Models in the Medical Domain?
Continual pre-training on a German medical corpus lets 7B models close much of the performance gap with 24B general models on medical benchmarks, though merging introduces some language mixing and verbosity.
-
FedProxy: Federated Fine-Tuning of LLMs via Proxy SLMs and Heterogeneity-Aware Fusion
FedProxy replaces weak adapters with a proxy SLM for federated LLM fine-tuning, outperforming prior methods and approaching centralized performance via compression, heterogeneity-aware aggregation, and training-free fusion.
-
MASS-RAG: Multi-Agent Synthesis Retrieval-Augmented Generation
MASS-RAG uses distinct agents for evidence summarization, extraction, and reasoning, then synthesizes their outputs to improve answer quality over standard RAG baselines on four benchmarks, especially when evidence is distributed.
-
Understanding Secret Leakage Risks in Code LLMs: A Tokenization Perspective
BPE tokenization creates gibberish bias in CLLMs, causing secrets with high character entropy but low token entropy to be preferentially memorized due to training data distribution shifts.
-
ALDEN: Boosting Private Data Extraction from Retrieval-Augmented Generation Systems via Active Learning and Distribution Estimation
ALDEN boosts private data extraction rates from RAG systems by combining active learning for query diversification with dynamic estimation of the underlying knowledge-base topic distribution.
-
Enhancing Target-Guided Proactive Dialogue Systems via Conversational Scenario Modeling and Intent-Keyword Bridging
Conversational scenario modeling from user profiles and domain knowledge, combined with intent-keyword bridging, improves proactivity, fluency, and informativeness in target-guided proactive dialogue systems.
-
From Traditional Taggers to LLMs: A Comparative Study of POS Tagging for Medieval Romance Languages
LLM-based POS tagging outperforms traditional taggers on medieval Occitan, Catalan, and French, with fine-tuning and cross-lingual transfer providing the largest gains for under-resourced varieties.
-
Gemma: Open Models Based on Gemini Research and Technology
Gemma introduces open 2B and 7B LLMs derived from Gemini technology that beat comparable open models on 11 of 18 text tasks and come with safety assessments.
-
A Case-Driven Multi-Agent Framework for E-Commerce Search Relevance
A case-driven multi-agent system automates the full pipeline of bad-case detection, annotation, and resolution for e-commerce search relevance using Annotator, Optimizer, and User agents plus supporting components.
-
Fine-Tuning Pre-Trained Code Models for AI-Generated Code Detection
Fine-tuning CodeBERT, GraphCodeBERT, UniXcoder and CodeT5+ with augmentation, cross-validation and ensembling yields macro-F1 of 0.737 on binary human-vs-AI code detection and 0.422 on 11-class model attribution in SemEval-2026 Task 13.
-
Gemma 2: Improving Open Language Models at a Practical Size
Gemma 2 models achieve leading performance at their sizes by combining established Transformer modifications with knowledge distillation for the 2B and 9B variants.
-
A Survey on Knowledge Distillation of Large Language Models
A comprehensive survey of knowledge distillation for LLMs structured around algorithms, skill enhancement, and vertical applications, highlighting data augmentation as a key enabler.
- Lessons from the Trenches on Reproducible Evaluation of Language Models