United States

While Large Reasoning Models (LRMs) excel at complex tasks via long Chain-of-Thought (CoT) reasoning, their outputs are often excessively verbose, leading to inefficiency. This problem is amplified when the student’s long-form reasoning mismatches the concise outputs of smaller teacher models—common in LLM distillation to avoid using costly large teachers. To address this issue, we propose Distilled Reasoning Pruning (DRP), a hybrid framework that combines inference-time pruning with tuning-based distillation. DRP leverages a teacher model to perform mathematical problem-solving skill-aware step decomposition and pruning, then distills the refined reasoning paths into a student model, enabling efficient and accurate reasoning. Across challenging math datasets, DRP significantly reduces token usage without sacrificing accuracy—for instance, cutting tokens on GSM8K from 917 to 328 while improving accuracy from 91.7% to 94.1%, and reducing AIME tokens by 43% with no performance drop. Further analysis shows that aligning training CoT structure with the student’s capacity is key to effective knowledge transfer.

ACL 2026

 DRP: Distilled Reasoning Pruning with Mathematical Skill-aware Step Decomposition for Efficient Large Reasoning Models

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community." 

*Message from the General Chair, Dirk Hovy*

EMNLP 2025

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

poster

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. 
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) 
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

Registration Required

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

Large language models (LLMs) rely on both contextual knowledge and parametric memory, yet these sources can conflict. Prior analysis largely focused on contextual question answering, suggesting that models tend to favor parametric knowledge under conflict, but this setting assumes that tasks should always rely on the provided passage. It therefore remains unclear how LLMs behave when \emph{tasks demand different kinds and degrees of knowledge utilization}. We address this gap with a model-agnostic diagnostic framework that holds underlying knowledge constant while injecting controlled conflicts across tasks with varying knowledge requirements. Evaluating representative open-source LLMs, we find that: (1) performance degradation under conflict correlates with a task’s knowledge reliance rather than conflict plausibility alone; (2) strategies such as explanatory rationales or reiteration increase context reliance, helping context-only tasks but harming those that require parametric knowledge; and (3) these behaviors bias model-based evaluation, raising concerns about the reliability of LLMs as judges. Together, our findings show that context–memory conflict is fundamentally task-dependent and motivate task-aware approaches to balancing context and memory in LLM deployment and evaluation.

Task Matters: Knowledge Requirements Shape LLM Responses to Context–Memory Conflict

Role-playing prompts effectively steer Large Language Models (LLMs), yet the neural mechanism driving this behavioral shift remains unclear. In this work, we identify Role-Sensitive Neurons (RSNs)—a sparse sub-network (≈ 0.5% of all neurons) governing the transition from hesitation to action. Using a novel evaluation framework with explicit abstention (MMLU-E), we reveal a Confidence-Performance Decoupling: roles primarily modulate the model's probabilistic "willingness to act" rather than its underlying knowledge representation. We demonstrate that RSNs function as a mechanistic gain control system: causal intervention on this subspace allows precise regulation of abstention behavior. Furthermore, cross-model transfer experiments confirm that these circuits are indigenous to pre-training, with Instruction Tuning (SFT) acting merely as a "signal sharpener" to refine latent gain dynamics. Finally, we identify a critical safety boundary: in knowledge-deficient models, amplifying RSNs induces "unwarranted certainty," highlighting decisiveness as a tunable gain parameter distinct from epistemic truth.

Role-Sensitive Neurons: A Neuron-Level Gain Control Mechanism for Confidence Steering

While Chain-of-Thought (CoT) reasoning enhances code generation in Large Language Models (LLMs), it introduces a critical challenge in uncertainty estimation: Confidence Saturation. Existing calibration methods, such as Self-Consistency, rely on the assumption that consensus implies correctness. This assumption fails under systematic errors, where models confidently repeat flawed logic, leading to miscalibrated high-confidence predictions. To address this, we introduce NeuroSym-Cal, a hierarchical calibration framework. We posit that reliable confidence requires interrogating the model at two complementary levels: the extrinsic consensus of its symbolic outputs and the intrinsic sensitivity of its latent reasoning. Specifically, we propose Reasoning Sensitivity Analysis to measure the local curvature of the deductive process via latent perturbation, providing a fine-grained signal that persists even when output consensus saturates. These orthogonal features are fused by a Contextual Calibration Network to predict correctness. Experiments across state-of-the-art reasoning models (e.g., DeepSeek-R1) demonstrate that NeuroSym-Cal effectively de-saturates overconfident errors, achieving state-of-the-art Expected Calibration Error (ECE) and superior selective generation performance on Out-Of-Domain (OOD) benchmarks.

NeuroSym-Cal: Bridging the Reasoning-Execution Gap in Code Generation via Hierarchical Calibration

Although negation is known to challenge large language models (LLMs), benchmarks for evaluating negation understanding—especially in Korean—are scarce. We conduct a corpus-based analysis of Korean negation and show that LLM performance degrades under negation. We then introduce *Thunder-KoNUBench*, a sentence-level negation understanding benchmark that reflects the empirical distribution of Korean negation phenomena. Evaluating 47 LLMs on Thunder-KoNUBench, we analyze the effects of model size and instruction tuning, and perform error analysis to better understand model behavior. We further show that fine-tuning on Thunder-KoNUBench improves negation understanding and broader contextual comprehension in Korean.

 Thunder-KoNUBench: A Corpus-Aligned Benchmark for Korean Negation Understanding

Text-to-Video (T2V) generation has benefited from recent advances in diffusion models, yet current systems still struggle under complex scenarios, which are generally exacerbated by the ambiguity and underspecification of text prompts. In this work, we formulate complex-scenario prompt refinement as a stage-wise multi-agent refinement process and propose SCMAPR, i.e., a scenario-aware and Self-Correcting Multi-Agent Prompt Refinement framework for T2V prompting. SCMAPR coordinates specialized agents to (i) route each prompt to a taxonomy-grounded scenario for strategy selection, (ii) synthesize scenario-aware rewriting policies and perform policy-conditioned refinement, and (iii) conduct structured semantic verification that triggers conditional revision when violations are detected. To clarify what constitutes complex scenarios in T2V prompting, provide representative examples, and enable rigorous evaluation under such challenging conditions, we further introduce T2V-Complexity, which is a complex-scenario T2V benchmark consisting exclusively of complex-scenario prompts. Extensive experiments on 3 existing benchmarks and our T2V-Complexity benchmark demonstrate that SCMAPR consistently improves text-video alignment and overall generation quality under complex scenarios, achieving up to 2.67% and 3.28 gains in average score on VBench and EvalCrafter, and up to 0.028 improvement on T2V-CompBench over 3 State-Of-The-Art baselines. The codes of SCMAPR are publicly available at https://github.com/HiThink-Research/SCMAPR.

SCMAPR: Self-Correcting Multi-Agent Prompt Refinement for Complex-Scenario Text-to-Video Generation

Large Language Models (LLMs) promise impressive capabilities, yet their multi-billion parameter scale makes on-device or low-resource deployment prohibitive. Mixed precision quantization offers a compelling solution, but existing methods struggle when the average precision drops below four bits, as they rely on isolated, layer-specific metrics that overlook critical inter-layer interactions affecting overall performance. To address these limitations, we first frame the mixed-precision quantization problem as a cooperative game among layers and introduce Shapley-based Progressive Quantization Estimation (SPQE) to efficiently obtain accurate Shapley estimates of layer sensitivities and inter-layer interactions. Leveraging the SPQE estimates, we propose Cooperative Game Inspired Mixed-Precision Quantization (CoopQ) which translates these Shapley estimates into a binary quadratic optimization formulation, assigning either 2 or 4-bit precision to layers under strict memory constraints. Comprehensive experiments conducted on Llama-3, Gemma-2, and Qwen models across three independent PTQ backends (Quanto, HQQ, GPTQ) demonstrate CoopQ’s scalability and consistently superior performance compared to methods relying solely on isolated metrics. Across average precisions spanning 4 bit down to 2 bit, CoopQ cuts Perplexity by 20 – 80 % relative to the best baseline, with the margin growing as the bit-width tightens.

CoopQ: Cooperative Game Inspired Layerwise Mixed Precision Quantization for LLMs

Automated scoring models are increasingly used to assign rubric-based quality ratings to complex language performances, including classroom transcripts, yet they typically provide little insight into why a particular score is produced. We propose a general framework for sentence-level interpretability of rubric-based scoring that combines model-agnostic Shapley-value attributions with rationales generated by large language models (LLMs). Instantiated on the Quality of Feedback dimension of the CLASS framework using the NCTE corpus, the framework enables systematic comparison of fine-tuned pretrained language models (PLMs) and prompted LLMs on both scoring performance and explanation faithfulness. Across 6k annotated transcript segments, fine-tuned PLMs outperform LLMs in prediction accuracy but exhibit label compression toward mid-scale scores. Deletion-based tests show that SHAP identifies sentences that reliably drive model predictions, producing typically larger and more coherent prediction shifts than LLM-generated rationales. Cross-model analyses further reveal that SHAP attributions transfer robustly across architectures, whereas LLM rationales exert limited and inconsistent influence. Overall, the findings demonstrate that SHAP provides more faithful and transferable explanations for rubric-based scoring, and that the proposed framework offers a principled basis for evaluating both scoring models and their explanations in high-stakes educational settings and other rubric-based language assessment tasks.

From Scoring to Explanations: Evaluating SHAP and LLM Rationales for Rubric-based Teaching Quality Assessment

Artificial intelligence (AI) is rapidly transforming healthcare, enabling the fast development of tools such as stress monitors, wellness trackers, and mental health chatbots. However, this rapid and low-barrier development can also introduce risks, including bias, privacy violations, and unequal access, especially when systems overlook real-world contexts, diverse user needs, and cultural settings. Many recent approaches use AI to identify such risks automatically, but this can reduce human engagement in understanding how harms arise, who they affect, and which stakeholder needs remain unspoken. We present a human-centered ethical foresight framework that generates speculative user stories and supports multi-agent discussions to help people reflect on potential benefits and harms of healthcare AI before deployment. In a user study, participants who engaged with stories identified a broader range of harms, distributing their responses more evenly across all 17 harm types, whereas those who did not engage with stories focused primarily on privacy and well-being (79.1%). Overall, our findings suggest that storytelling helps people anticipate potential risks and benefits and reflect more broadly on how AI systems may affect different users, contexts, and often unspoken needs.

Telling Speculative Stories to Help Humans Imagine the Harms of Healthcare AI

We study the reasoning behavior of large language models (LLMs) under limited computation budgets. In such settings, producing useful partial solutions quickly is often more practical than exhaustive reasoning, which incurs high inference costs. Many real-world tasks, such as trip planning, require models to deliver the best possible output within a fixed reasoning budget. We introduce an anytime reasoning framework and the Anytime Index, a metric that quantifies how effectively solution quality improves as reasoning tokens increase. To further enhance efficiency, we propose an inference-time self-improvement method using LLM-synthesized preference data, where models learn from their own reasoning comparisons to produce better intermediate solutions. Experiments on NaturalPlan (Trip), AIME, and GPQA datasets show consistent gains across Grok-3, GPT-oss, GPT-4.1/4o, and LLaMA models, improving both reasoning quality and efficiency under budget constraints.

 Budget-Aware Anytime Reasoning with LLM-Synthesized Preference Data

While Retrieval-Augmented Generation(RAG) enhances multi-modal large language models(MLLMs) by introducing external knowledge, existing RAG systems still face significant limitations when dealing with complex visual reasoning. On one hand, MLLMs, being generative models, produce suboptimal embeddings for retrieval tasks. On the other hand, existing methods naively insert images into context without adequate visual perception, thereby limiting reasoning capabilities. To address these challenges, we propose MDocRAG-RL, a novel RAG framework for complex visual reasoning. We design specialized pre-training and fine-tuning tasks to enable MLLMs to compress visual document representations and align textual and visual embeddings for improved retrieval efficiency. Additionally, we design a visual perception action space for the generator that allows progressive coarse-to-fine information acquisition from visually-rich documents. Furthermore, we develop a reinforcement learning framework to enhance the complex visual reasoning capability of the RAG system. Extensive experiments on multiple challenging benchmarks demonstrate the significant effectiveness of our approach, achieving state-of-the-art performance across various benchmarks.

Downloads

Next from ACL 2026

Task Matters: Knowledge Requirements Shape LLM Responses to Context–Memory Conflict

Similar lecture

Improving Reasoning Capabilities in Small Models through Mixture-of-layers Distillation with Stepwise Attention on Key Information