United States

Large language models (LLMs) trained on unfiltered corpora inherently risk retaining sensitive information, necessitating selective knowledge unlearning for regulatory compliance and ethical safety. However, existing parameter-modifying methods face fundamental limitations: high computational costs, uncontrollable forgetting boundaries, and strict dependency on model weight access. These constraints render them impractical for closed-source models, yet current non-invasive alternatives remain unsystematic and reliant on empirical experience. To address these challenges, we propose the Controllable Alignment Prompting for Unlearning (CAP) framework, an end-to-end prompt-driven unlearning paradigm. CAP decouples unlearning into a learnable prompt optimization process via reinforcement learning, where a prompt generator collaborates with the LLM to suppress target knowledge while preserving general capabilities selectively. This approach enables reversible knowledge restoration through prompt revocation. Extensive experiments demonstrate that CAP achieves precise, controllable unlearning without updating model parameters, establishing a dynamic alignment mechanism that overcomes the transferability limitations of prior methods.

ACL 2026

CAP: Controllable Alignment Prompting for Unlearning in LLMs

## 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.

technical paper

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.

**Online Registration Form**: https://acl.swoogo.com/acl2026

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.

Proactive dialogue has emerged as a critical and challenging research problem in advancing large language models (LLMs). Existing works predominantly focus on domain-specific or task-oriented scenarios, which leads to fragmented evaluations and limits the comprehensive exploration of models' proactive dialogue abilities. In this work, we propose ProactiveEval, a unified framework for evaluating proactive dialogue capabilities of LLMs. This framework decomposes proactive dialogue into target planning and dialogue guidance, establishing evaluation metrics across various domains. Moreover, it also enables the automatic generation of diverse and challenging evaluation data. Based on the proposed framework, we develop 328 evaluation environments spanning 6 distinct domains. Through experiments with 22 different types of LLMs, we show that DeepSeek-R1 and Claude-3.7-Sonnet exhibit exceptional performance on target planning and dialogue guidance tasks, respectively. Finally, we investigate how reasoning capabilities influence proactive behaviors and discuss their implications for future model development. Our code and data are available at the https://github.com/liutj9/ProactiveEval.

ProactiveEval: A Unified Evaluation Framework for Proactive Dialogue Agents

Existing Reinforcement Learning with Verifiable Rewards (RLVR) algorithms, such as GRPO, rely on rigid, uniform, and symmetric trust region mechanisms that are fundamentally misaligned with the complex optimization dynamics of Large Language Models (LLMs). In this paper, we identify three critical challenges in these methods: (1) inefficient gradient utilization caused by the binary cutoff of hard clipping, (2) insensitive probability mass arising from uniform ratio constraints that ignore the token distribution, and (3) asymmetric signal reliability stemming from the disparate credit assignment ambiguity between positive and negative samples. To bridge these gaps, we propose Mass-Adaptive Soft Policy Optimization (MASPO), a unified framework designed to harmonize these three dimensions. MASPO integrates a differentiable soft Gaussian gating to maximize gradient utility, a mass-adaptive limiter to balance exploration across the probability spectrum, and an asymmetric risk controller to align update magnitudes with signal confidence. Extensive evaluations demonstrate that MASPO serves as a robust, all-in-one RLVR solution, significantly outperforming baselines. Our code is available at: https://github.com/FlyTune/MASPO-RL.

MASPO: Unifying Gradient Utilization, Probability Mass, and Signal Reliability for Robust and Sample-Efficient LLM Reasoning

Existing \textit{In-context Learning (ICL)} typically assumes the retrieval dataset contains demonstrations for all output label spaces. However, in real-world scenarios, delays in dataset updates or incomplete data annotation may result in the retrieval dataset containing labeled demonstrations for only a subset of the output space. We refer to this phenomenon as an \textit{incomplete retrieval dataset} and define the in-context learning under this condition as \textit{Incomplete In-context Learning (IICL)}. To address IICL, we propose \textit{Iterative Judgments and Integrated Prediction (IJIP)}, a framework with train-free and train-based variants. For classification, the iterative judgments stage of IJIP reformulates an (m)-class problem into (m) binary tasks, converting IICL into standard ICL. The integrated prediction stage of IJIP then refines results using both the input and initial predictions. We further extend IJIP to text regression and generation, and introduce lightweight variants that reduce computation and token costs. Across six LLMs, seven tasks, and eight datasets, IJIP achieves state-of-the-art results under two incompleteness settings and even outperforms standard ICL with complete labels. IJIP also supports a semi-supervised variant and can serve as a plug-and-play enhancement for existing ICL and zero-shot methods.

Incomplete In-context Learning

Vision-Language Models (VLMs) have demonstrated impressive capabilities in code generation across various domains. However, their ability to replicate complex, multi-panel visualizations from real-world data remains largely unassessed. To address this gap, we introduce \textbf{\texttt{RealChart2Code}}, a new large-scale benchmark with over 2,800 instances grounded in authentic datasets and featuring tasks with clear analytical intent. Crucially, it is the first benchmark to systematically evaluate chart generation from large-scale raw data and assess iterative code refinement in a multi-turn conversational setting. Our comprehensive evaluation of 14 leading VLMs on \texttt{RealChart2Code} reveals significant performance degradation compared to simpler benchmarks, highlighting their struggles with complex plot structures and authentic data. Our analysis uncovers a substantial performance gap between proprietary and open-weight models and confirms that even state-of-the-art VLMs often fail to accurately replicate intricate, multi-panel charts. These findings provide valuable insights into the current limitations of VLMs and guide future research directions.

RealChart2Code: Bridging the Gap in Real-World Chart-to-Code Generation via Multi-Task Evaluation

Math word problems (MWPs) are critical K-12 educational tools, and customizing them to students' interests and ability levels can enhance learning. However, teachers struggle to find time to customize MWPs for students given large class sizes and increasing burnout. We propose that LLMs can support math education by generating MWPs customized to student interests and math education standards. We use a joint human expert-LLM judge approach to evaluate over 11,000 MWPs generated by open and closed LLMs and develop the first teacher-annotated dataset for standards-aligned educational MWP generation. We show the value of our data by using it to train a 12B open model that matches the performance of larger and more capable open models. We also use our teacher-annotated data to train a text classifier that enables a 30B open LLM to outperform existing closed baselines without any training. Next, we show our models' MWPs are more similar to human-written MWPs than those from existing models. We conclude by conducting the first study of customized LLM-generated MWPs with grade school students, finding they perform similarly on our models' MWPs relative to human-written MWPs but consistently prefer our customized MWPs.

EDUMATH: Generating Standards-aligned Educational Math Word Problems

Although Large Language Models (LLMs) have demonstrated substantial proficiency in reasoning, current approaches focus disproportionately on scaling correct training samples, underexploring the value of incorrect reasoning trajectories. Motivated by how humans learn from mistakes, we propose ReActR (Reasoning through Error-Activated Reflection), a framework that enhances reasoning by learning reflective behaviors from erroneous trajectories. Specifically, ReActR comprises data construction and training. First, we synthesize multi-turn erroneous reasoning dataset spanning diverse error types and difficult levels via self-generation and targeted error generation. Second, we enhance the model's capabilities through Supervised Fine-Tuning (SFT) on synthesized data and then apply Group Relative Policy Optimization (GRPO) with multiple reward signals to further refine reasoning performance. Extensive experiments across five benchmarks and three LLMs demonstrate that ReActR effectively enhances reasoning performance. Notably, on Llama-3-8B, ReActR achieves an average improvement of 3.5% across the five datasets.

ReActR: Reasoning through Error-Activated Reflection for LLM Post-Training

The linear memory growth of the KV cache poses a significant bottleneck for LLM inference in long-context tasks. Existing static compression methods often fail to preserve globally important information. Although recent dynamic retrieval approaches attempt to address this issue, they typically suffer from coarse-grained caching strategies and incur high I/O overhead. To overcome these limitations, we propose HeteroCache, a training-free dynamic compression framework. Our method is built on two key insights: attention heads exhibit diverse temporal heterogeneity, and there is significant spatial redundancy among heads within the same layer.
Guided by these insights, HeteroCache categorizes heads based on stability and similarity, applying a fine-grained weighting strategy that allocates larger cache budgets to heads with rapidly shifting attention to capture context changes.
Furthermore, it features a hierarchical storage mechanism where representative heads monitor attention drift to trigger asynchronous, on-demand context retrieval, thereby hiding I/O latency.
Experiments demonstrate that HeteroCache achieves state-of-the-art performance on long-context benchmarks and accelerates decoding by up to 3× compared to the original model with a 224K context. Our code is available at https://github.com/ponytaill/HeteroCache.

HeteroCache: A Dynamic Retrieval Approach to Heterogeneous KV Cache Compression for Long-Context LLM Inference

LLM-based agents depend on effective tool-use policies to solve complex tasks, yet optimizing these policies remains challenging due
to delayed supervision and the difficulty of
credit assignment in long-horizon trajectories.
Existing optimization approaches tend to be
either monolithic, which are prone to entangling behaviors, or single-aspect, which ignore
cross-module error propagation. To address
these limitations, we propose EVOTOOL, a self-
evolving framework that optimizes a modular
tool-use policy via a gradient-free evolutionary paradigm. EVOTOOL decomposes agent’s
tool-use policy into four modules, including
Planner, Selector, Caller, and Synthesizer, and
iteratively improves them through three mechanisms. Trajectory-Grounded Blame Attribution uses diagnostic traces to localize failures
to a specific module. Feedback-Guided Targeted Mutation then edits only that module
via natural-language critique. Diversity-Aware
Population Selection preserves complementary
candidates to ensure solution diversity. Across
four diverse benchmarks, EVOTOOL outperforms strong baselines by over 5 points on both
GPT-4.1 and Qwen3-8B, while achieving superior efficiency and transferability.

EVOTOOL: Self-Evolving Tool-Use Policy Optimization in LLM Agents via Blame-Aware Mutation and Diversity-Aware Selection

Reviewer assignment is increasingly critical yet challenging in the LLM era, where rapid topic shifts render many pre-2023 benchmarks outdated and where proxy signals poorly reflect true reviewer familiarity. We address this evaluation bottleneck by introducing LR-bench, a high-fidelity, up-to-date benchmark curated from 2024--2025 AI/NLP manuscripts with five-level self-assessed familiarity ratings collected via a large-scale email survey, yielding 1,055 expert-annotated paper--reviewer--score annotations. We further propose a reviewer-centric ranking framework that distills each reviewer’s recent publications into compact keyword-based profiles and fine-tunes an embedding model with weak preference supervision constructed from heuristic retrieval signals, enabling to match each manuscript against a reviewer profile directly. Across the LR-bench and the CMU gold-standard dataset, our approach consistently achieves state-of-the-art performance, outperforming strong embedding baselines by a clear margin. We release LR-bench at https://huggingface.co/datasets/Gnociew/LR-bench, and an github repository at https://github.com/Gnociew/RATE-Reviewer-Assignment.

RATE: Reviewer Profiling and Annotation-free Training for Expertise Ranking in Peer Review Systems

Large language models (LLMs) have demonstrated increasingly strong reasoning capabilities, achieving remarkable progress in knowledge graph question answering (KGQA). However, a key challenge in such systems is non-deterministic reasoning, where the model indecisively activates multiple semantically related knowledge graph edges for a given query, frequently leading to incorrect answers. To address this issue, we propose Diagnosing and Remedying Representation Deficiencies for Deterministic Reasoning in KGQA (DR2). DR2 identifies and localizes non-deterministic reasoning behaviors, uncovering the underlying semantic representation deficiencies in LLMs. Building on this diagnosis, we design abductive reasoning-based preference learning, which promotes fine-grained semantic discrimination and mitigates non-deterministic reasoning errors. Experimental results demonstrate that the proposed DR2 significantly outperforms several strong baselines, achieving state-of-the-art performance on the widely used WebQSP and CWQ benchmarks.

Downloads

Next from ACL 2026

ProactiveEval: A Unified Evaluation Framework for Proactive Dialogue Agents

Similar lecture

Does Localization Inform Unlearning? A Rigorous Examination of Local Parameter Attribution for Knowledge Unlearning in Language Models

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES