United States

Large language models (LLMs) demonstrate strong performance on multi-step reasoning tasks by producing intermediate explanations, commonly referred to as chains of thought (CoTs). However, the generated rationales are typically verbose, consuming many additional tokens, and thus degrading throughput and increasing inference energy consumption. Interestingly, we find that verbose and concise CoTs correspond to distinct regions in the model&#39;s intermediate activation space, suggesting that verbosity is a steerable latent attribute. Building on this observation, we develop an inference-time method to automatically steer the model response towards concise reasoning traces without updating model parameters. Our method, dubbed _ASC_ (Activation-Steered Compression), generates concise CoTs by directly adjusting internal representations via activation steering. A key component of ASC is **Contrastive Energy-Based Steering (CES)**, a principled procedure to learn a _single_ steering vector from a small set of verbose–concise CoT pairs by optimizing a length-normalized contrastive energy objective. To further ensure reliable steering and preserve general utility, CES enforces a differentiable **KL trust region** during steering vector optimization, explicitly constraining the distribution shift within a specified budget. With only 100 pairs of verbose–concise examples, ASC reduces the generated token length by as much as 69.4% across five reasoning benchmarks (MATH500, GSM8K, LiveCodeBench, GSM8K-Hard, and AQuA-RAT) while maintaining accuracy across models with 1.5B, 7B, 8B, and 32B parameters. On MATH500, ASC achieves an end-to-end inference speed-up of 2.7× on an 8B model.

ACL 2026

 Activation Steering for Chain-of-Thought Compression

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. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<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>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>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) demonstrate strong performance on multi-step reasoning tasks by producing intermediate explanations, commonly referred to as chains of thought (CoTs). However, the generated rationales are typically verbose, consuming many additional tokens, and thus degrading throughput and increasing inference energy consumption. Interestingly, we find that verbose and concise CoTs correspond to distinct regions in the model's intermediate activation space, suggesting that verbosity is a steerable latent attribute. Building on this observation, we develop an inference-time method to automatically steer the model response towards concise reasoning traces without updating model parameters. Our method, dubbed _ASC_ (Activation-Steered Compression), generates concise CoTs by directly adjusting internal representations via activation steering. A key component of ASC is **Contrastive Energy-Based Steering (CES)**, a principled procedure to learn a _single_ steering vector from a small set of verbose–concise CoT pairs by optimizing a length-normalized contrastive energy objective. To further ensure reliable steering and preserve general utility, CES enforces a differentiable **KL trust region** during steering vector optimization, explicitly constraining the distribution shift within a specified budget. With only 100 pairs of verbose–concise examples, ASC reduces the generated token length by as much as 69.4% across five reasoning benchmarks (MATH500, GSM8K, LiveCodeBench, GSM8K-Hard, and AQuA-RAT) while maintaining accuracy across models with 1.5B, 7B, 8B, and 32B parameters. On MATH500, ASC achieves an end-to-end inference speed-up of 2.7× on an 8B model.

poster

Translation-based prompting is widely used in multilingual LLMs, yet its effectiveness varies across languages and tasks. We evaluate prompting strategies across ten languages of different resource levels and four benchmarks. Our analysis shows that no single strategy is universally optimal. Translation strongly benefits low-resource languages even when translation quality is imperfect, high-resource languages gain little, and prompt-based self-routing underperforms explicit translation. Motivated by these findings, we formulate prompting strategy selection as a learned decision problem and introduce lightweight classifiers that predict whether native or translation-based prompting is optimal for each instance. The classifiers achieve statistically significant improvements over fixed strategies across four benchmarks and generalize to unseen task formats not observed during training. Further analysis reveals that language resource level, rather than translation quality alone, determines when translation is beneficial.

No One Fits All: From Fixed Prompting to Learned Routing in Multilingual LLMs

While Large Language Models (LLMs) demonstrate remarkable capabilities in open-ended creative generation, they notably struggle with Format-Constrained Generation tasks—such as poetry and lyrics—where strict adherence to multidimensional structural constraints (i.e., format, phonetics, and rhyme) is prerequisite to aesthetic value. Existing paradigms predominantly rely on unreliable prompting or rigid constrained decoding strategies; the former often fails to ensure compliance, while the latter compromises inference latency and disrupts the natural probability distribution, degrading generation quality. To bridge this gap, we establish CCP-Arena, a rigorous testbed for Chinese Classical Poetry, and proposeProgressive Structural Internalization (PSI)} a novel framework designed to embed external constraints into the model's intrinsic intuition. PSI initiates withStructural Scaffolding via Explicit Cognitive Planning, utilizing explicit template to provide a structural scaffold for subsequent generation. This is followed by a Cascaded Reinforcement Learning stage guided by a Holistic Reward Model, which optimizes for precise structural-semantic alignment. Extensive experiments demonstrate that PSI achieves state-of-the-art performance, surpassing baselines in both strict constraint adherence and literary aesthetics. Furthermore, mechanistic analysis confirms that our method effectively internalizes structural information into the model's latent representations, offering a robust and efficient solution for constrained creative generation.

From Scaffolding to Assimilation: Progressive Structural Internalization for Format-Constrained Creative Text Generation

Standardized benchmarks have become the dominant metric for measuring progress in large language models, yet their validity is increasingly compromised by data contamination and the unclear relationship between benchmark scores and genuine language understanding. We introduce Gaperon, a suite of fully open bilingual (French-English) language models designed as an experimental testbed to investigate evaluation dynamics under realistic training conditions. Our study makes three core contributions. First, we demonstrate mismatches between benchmark performance and generation quality: models that excel on benchmarks may underperform in qualitative text generation, and vice versa. Second, through our deliberately contaminated Gaperon-Garlic variant, we show that competitive benchmark scores can be recovered via late-stage contamination with only moderate degradation of generation quality, and surprisingly, such contamination also improves performance on held-out benchmarks. Third, we provide empirical evidence that widely used neural quality filters, particularly those trained to favor instructional or educational content, amplify benchmark contamination in pretraining corpora, with the DCLM classifier systematically ranking benchmark samples in the top-5 percentiles of samples. We release all models, data mixtures, checkpoints, and evaluation code to support reproducibility and further investigation.

Gaperon: A Peppered English-French Generative Language Model Suite

As large language model (LLM) based systems take on high-stakes roles in real-world decision-making, they must reconcile competing instructions from multiple sources within a single prompt context. Enforcing an instruction hierarchy, where higher-level directives override lower-priority requests, is critical to the reliability and controllability of LLMs. In this work, we reframe instruction hierarchy resolution as a reasoning task. The model must first "think" about the relationship between a given user prompt and higher-priority instructions before generating a response. To enable this capability, we construct VerIH, a training dataset of constraint-following tasks with verifiable answers, comprising aligned and conflicting system--user instructions. We show that lightweight reinforcement learning with VerIH effectively transfers general reasoning capabilities of models to instruction prioritization. Our method leads to consistent improvements across multiple model families and scales on instruction following and instruction hierarchy benchmarks, achieving 
~20% absolute improvement in conflict setups. Our method also generalizes to safety-critical scenarios beyond the training distribution, exhibiting increased robustness against jailbreak and prompt injection, reducing absolute attack success rates by up to 20%. Our results establish reasoning over instruction hierarchies as a practical mechanism for improving AI reliability, where targeted updates to system prompts produce predictable, controllable, and robust changes in model behavior.

Reasoning Up the Instruction Ladder for Controllable Language Models

As numerous instruction-tuning datasets continue to emerge, dynamically balancing and optimizing their mixtures has become a critical
challenge. To address this, we propose DynamixSFT, a dynamic and automated method for instruction-tuning dataset mixture optimization. We formulate the problem as a multi-armed bandit setup and introduce a Prior-scaled Boltzmann Exploration that softly anchors the updated sampling distribution to the original dataset proportions, thereby preserving the inherent diversity and coverage of the collection. Sampling probabilities are updated using a lightweight 1-Step Look-ahead Reward, reflecting how much the dataset contributes to improving the model’s performance at its current state. We demonstrate that DynamixSFT effectively optimizes the TÜLU-2-mixture and
TÜLU-3-mixture collections across 10 benchmarks, while introducing minimal computational overhead over naive sampling. Furthermore, we provide a comprehensive analysis and visualizations to offer deeper insights into the adaptive dynamics of our method.

DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning Collections

Large language models (LLMs) are primarily accessed via commercial APIs, but this often requires users to expose their data to service providers. In this paper, we explore how users can stay in control of their data by using privacy profiles: simple natural language instructions that say what should and should not be revealed. We build a framework where a local model uses these instructions to rewrite queries, only hiding details deemed sensitive by the user, before sending them to an external model, thus balancing privacy with performance. To support this research, we introduce PEEP, a multilingual dataset of real user queries annotated to mark private content and paired with synthetic privacy profiles, alongside PROFIT, a training procedure that enables effective and efficient use of the pipeline. Experiments with lightweight local LLMs show that, after training, they not only achieve markedly better privacy preservation but also match or exceed the performance of much larger few-shot models.

Controlling What You Share: Assessing Language Model Adherence to Privacy Preferences

Answering open-ended questions remains challenging for AI systems because it requires synthesis, judgment, and exploration beyond factual retrieval, and users often refine answers through multiple iterations rather than accepting a single response. Existing QA benchmarks do not explicitly support this refinement process. To address this gap, we introduce a new task, document-grounded related insight generation, where the goal is to generate additional insights from a document collection that help improve, extend, or rethink an initial answer to an open-ended question, ultimately supporting richer user interaction and a better overall question answering experience. We curate and release SCOpE-QA (Scientific Collections for Open-Ended QA), a dataset of 3,000 open-ended questions across 20 research collections. We present InsightGen, a two-stage approach that first constructs a thematic representation of the document collection using clustering, and then selects related context based on neighborhood selection from the thematic graph to generate diverse and relevant insights using LLMs. Extensive evaluation on 3,000 questions using two generation models and two evaluation settings shows that InsightGen consistently produces useful, relevant, and actionable insights, establishing a strong baseline for this new task.

An Answer is just the Start: Related Insight Generation for Open-Ended Document-Grounded QA

LLM-powered coding agents are redefining how real-world software is developed. To drive the research towards better coding agents, we require challenging benchmarks that can rigorously evaluate the ability of such agents to perform various software engineering tasks. However, popular coding benchmarks such as HumanEval and SWE-Bench focus on narrowly scoped tasks such as competition programming and patch generation. In reality, software engineers have to handle a broader set of tasks for real-world software development. To address this gap, we propose OmniCode, a novel software engineering benchmark that contains a broader and more diverse set of task categories beyond code or patch generation. Overall, OmniCode contains 1794 tasks spanning three programming languages -- Python, Java, and C++ -- and four key categories: bug fixing, test generation, code review fixing, and style fixing. In contrast to prior software engineering benchmarks, the tasks in OmniCode are (1) manually validated to eliminate ill-defined problems, and (2) synthetically crafted or recently curated to avoid data leakage issues, presenting a new framework for synthetically generating diverse software tasks from limited real-world data. We evaluate OmniCode with popular agent frameworks such as SWE-Agent and show that while they may perform well on bug fixing for Python, they fall short on tasks such as Test Generation and in languages such as C++ and Java. For instance, SWE-Agent achieves a maximum of 25.0% with DeepSeek-V3.1 on C++ Test Generation. OmniCode aims to serve as a robust benchmark and spur the development of agents that can perform well across different aspects of software development.

OmniCode: A Benchmark for Evaluating Software Development Agents

As large language models are increasingly adopted and trusted in real-world applications, understanding their behavior beyond single-turn prompting has become critical. Existing safety evaluations primarily focus on refusal-based methods that test whether models avoid responding to inappropriate or violent requests, leaving open questions about how models behave in interactive social settings. In this paper, we observe the adversarial behavior of LLM models through a multi-agent simulation across five diverse social deduction conversational games, acting as testbeds that provide social pressures and survival stress based on game design without explicit prompt injections. From these interactions, we construct a closed behavioral taxonomy derived through open card sorting, applied uniformly across models using a meta-LLM for behavior labeling. This approach displays that models exhibit distinct behavioral profiles and that models' different ways of structured deliberation influence both social stability and competitive success.

Measuring Large Language Models' Adversarial Behavior in Social Deduction Games

While Large Language Models (LLMs) show remarkable capabilities, their unreliability remains a critical barrier to deployment in high-stakes domains. This survey charts a functional evolution in addressing this challenge: the evolution of uncertainty from a passive diagnostic metric to an active control signal guiding real-time model behavior. We demonstrate how uncertainty is leveraged as an active control signal across three frontiers: in \textbf{advanced reasoning} to optimize computation and trigger self-correction; in \textbf{autonomous agents} to govern metacognitive decisions about tool use and information seeking; and in \textbf{reinforcement learning} to mitigate reward hacking and enable self-improvement via intrinsic rewards. By grounding these advancements in emerging theoretical frameworks like Bayesian methods and Conformal Prediction, we provide a unified perspective on this transformative trend. This survey provides a comprehensive overview, critical analysis, and practical design patterns, arguing that mastering the new trend of uncertainty is essential for building the next generation of scalable, reliable, and trustworthy AI.

Downloads

Next from ACL 2026

No One Fits All: From Fixed Prompting to Learned Routing in Multilingual LLMs

Similar lecture

Compositional Steering of Large Language Models with Steering Tokens

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES