Austria

Language models (LMs) can make a correct prediction based on many possible signals in a prompt, not all corresponding to recall of factual associations. However, current interpretations of LMs fail to take this into account. For example, given the query &quot;Astrid Lindgren was born in&quot; with the corresponding completion &quot;Sweden&quot;, no difference is made between whether the prediction was based on knowing where the author was born or assuming that a person with a Swedish-sounding name was born in Sweden. In this paper, we present a model-specific recipe - PrISM - for constructing datasets with examples of four different prediction scenarios: generic language modeling, guesswork, heuristics recall and exact fact recall. We apply two popular interpretability methods to the scenarios: causal tracing (CT) and information flow analysis. We find that both yield distinct results for each scenario. Results for exact fact recall and generic language modeling scenarios confirm previous conclusions about the importance of mid-range MLP sublayers for fact recall, while results for guesswork and heuristics indicate a critical role of late last token position MLP sublayers. In summary, we contribute resources for a more extensive and granular study of fact completion in LMs, together with analyses that provide a more nuanced understanding of how LMs process fact-related queries.

ACL 2025

Fact Recall, Heuristics or Pure Guesswork? Precise Interpretations of Language Models for Fact Completion

data artifacts

knowledge tracing

probing

robustness

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s 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.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

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 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Language models (LMs) can make a correct prediction based on many possible signals in a prompt, not all corresponding to recall of factual associations. However, current interpretations of LMs fail to take this into account. For example, given the query "Astrid Lindgren was born in" with the corresponding completion "Sweden", no difference is made between whether the prediction was based on knowing where the author was born or assuming that a person with a Swedish-sounding name was born in Sweden. In this paper, we present a model-specific recipe - PrISM - for constructing datasets with examples of four different prediction scenarios: generic language modeling, guesswork, heuristics recall and exact fact recall. We apply two popular interpretability methods to the scenarios: causal tracing (CT) and information flow analysis. We find that both yield distinct results for each scenario. Results for exact fact recall and generic language modeling scenarios confirm previous conclusions about the importance of mid-range MLP sublayers for fact recall, while results for guesswork and heuristics indicate a critical role of late last token position MLP sublayers. In summary, we contribute resources for a more extensive and granular study of fact completion in LMs, together with analyses that provide a more nuanced understanding of how LMs process fact-related queries.

poster

Despite exceptional capabilities in knowledge-intensive tasks, Large Language Models (LLMs) face a critical gap in understanding how they internalize new knowledge, particularly how acquired knowledge becomes structurally embedded in their neural computations. We address this issue through the lens of knowledge circuit evolution, identifying computational subgraphs that facilitate knowledge storage and processing. Our systematic analysis of circuit evolution throughout continual pre-training reveals several key findings: (1) the acquisition of new knowledge is influenced by its relevance to pre-existing knowledge; (2) the evolution of knowledge circuits exhibits a distinct phase shift from formation to optimization; (3) the evolution of knowledge circuits follows a deep-to-shallow pattern. These insights not only advance our theoretical understanding of the mechanisms of new knowledge acquisition in LLMs, but also provide potential implications for improving continual pre-training strategies to enhance model performance.

How Do LLMs Acquire New Knowledge? A Knowledge Circuits Perspective on Continual Pre-Training

Temporal tabular question answering presents a significant challenge for Large Language Models (LLMs), requiring robust reasoning over structured data—a task where traditional prompting methods often fall short. These methods face challenges such as memorization, sensitivity to table size, and reduced performance on complex queries. To overcome these limitations, we introduce TEMPTABQA-C, a synthetic dataset designed for systematic and controlled evaluations, alongside a symbolic intermediate representation that transforms tables into database schemas. This structured approach allows LLMs to generate and execute SQL queries, enhancing generalization and mitigating biases. By incorporating adaptive fewshot prompting with contextually tailored examples, our method achieves superior robustness, scalability, and performance. Experimental results consistently highlight improvements across key challenges, setting a new benchmark for robust temporal reasoning with LLMs.

LLM-Symbolic Integration for Robust Temporal Tabular Reasoning

Large language models (LLMs) have transformed AI across diverse domains, with prompting being central to their success in guiding model outputs. However, manual prompt engineering is both labor-intensive and domain-specific, necessitating the need for automated solutions. We introduce PromptWizard, a novel, fully automated framework for discrete prompt optimization, utilizing a self-evolving, self-adapting mechanism. Through a feedback-driven critique and synthesis process, PromptWizard achieves an effective balance between exploration and exploitation, iteratively refining both prompt instructions and in-context examples to generate human-readable, task-specific prompts. This guided approach systematically improves prompt quality, resulting in superior performance across 45 tasks. PromptWizard excels even with limited training data, smaller LLMs, and various LLM architectures. Additionally, our cost analysis reveals a substantial reduction in API calls, token usage, and overall cost, demonstrating PromptWizard's efficiency, scalability, and advantages over existing prompt optimization strategies.

PromptWizard: Optimizing Prompts via Task-Aware, Feedback-Driven Self-Evolution

Multiple-choice exam questions with "None of the above" (NA) options have been extensively studied in educational testing, in which existing research suggests that they better assess true knowledge. However, their impact on Large Language Models (LLMs) evaluation remains underexplored. Through systematic experiments with 28 LLMs on the MMLU benchmark, we examine how NA options affect model performance and confidence calibration. Our analysis reveals that NA options, when used as the correct answer, lead to a consistent 30-50% performance drop across models regardless of scale--suggesting that LLMs lack the meta-cognitive ability to systematically evaluate and reject all given options when none are correct. This degradation shows strong domain dependence, with minimal impact on mathematical reasoning (14.6% drop) but severe effects on tasks requiring uncertainty handling like business ethics (48.1% drop). Our results highlight important implications for benchmark design and raise questions about LLMs' ability to handle uncertainty in real-world applications.

None of the Above, Less of the Right Parallel Patterns in Human and LLM Performance on Multi-Choice Questions Answering

Recent advances in handling long sequences have unlocked new possibilities for long-context in-context learning (ICL). While existing research predominantly focuses on performance gains driven by additional in-context examples, the impact on the trustworthiness of generated responses remains underexplored. This paper addresses this gap by investigating how increased examples influence predictive uncertainty—an essential aspect in trustworthiness. We begin by systematically quantifying uncertainty across different “shot” configurations in ICL, emphasizing the role of example quantity. Through uncertainty decomposition, we introduce a novel perspective on performance enhancement, focusing on epistemic uncertainty (EU). Our results reveal that additional examples reduce total uncertainty in both simple and complex tasks by injecting task-specific knowledge, thereby diminishing EU and enhancing performance. For complex tasks, these advantages emerge only after addressing the increased noise and uncertainty associated with longer inputs. Finally, we investigate the progression of internal confidence across layers, uncovering the underlying mechanisms that drive the reduction in uncertainty.

Uncertainty Unveiled: Can Exposure to More In-context Examples Mitigate Uncertainty for Large Language Models?

Audio-Visual Speech Recognition (AVSR) achieves robust speech recognition in noisy environments by combining auditory and visual information. However, recent Large Language Model (LLM) based AVSR systems incur high computational costs due to the high temporal resolution of audio-visual speech processed by LLMs. In this work, we introduce an efficient multimodal speech LLM framework that minimizes token length while preserving essential linguistic content. Our approach employs an early av-fusion module for streamlined feature integration, an audio-visual speech Q-Former that dynamically allocates tokens based on input duration, and a refined query allocation strategy with a speech rate predictor to adjust token allocation according to speaking speed of each audio sample. Extensive experiments on the LRS3 dataset show that our method achieves state-of-the-art performance with a WER of 0.74% while using only 2.8 tokens per second. Moreover, our approach not only reduces token usage by 88.8% compared to the previous multimodal speech LLM framework, but also improves computational efficiency by reducing FLOPs by 36%.

MMS-LLaMA: Efficient LLM-based Audio-Visual Speech Recognition with Minimal Multimodal Speech Tokens

Large Language Models (LLMs) rely on generating extensive intermediate reasoning units (e.g., tokens, sentences) to enhance final answer quality across a wide range of complex tasks. While this approach has proven effective, it inevitably increases substantial inference costs. Previous methods adopting token-level reduction without clear criteria result in poor performance compared to models trained with complete rationale. To address this challenge, we propose a novel sentence-level rationale reduction framework leveraging likelihood-based criteria, *verbosity*, to identify and remove redundant reasoning sentences. Unlike previous approaches, our method leverages *verbosity* to selectively remove redundant reasoning sentences while preserving reasoning capabilities. Our experimental results across various reasoning tasks demonstrate that our method improves performance by an average of 7.71% while reducing token generation by 19.87% compared to model trained with complete reasoning paths.

Verbosity-Aware Rationale Reduction: Sentence-Level Rationale Reduction for Efficient and Effective Reasoning

Retrieval-Augmented Generation (RAG) is a powerful strategy for improving the factual accuracy of models by retrieving external knowledge relevant to queries and incorporating it into the generation process. However, existing approaches primarily focus on text, with some recent advancements considering images, and they largely overlook videos, a rich source of multimodal knowledge capable of representing contextual details more effectively than any other modality. While very recent studies explore the use of videos in response generation, they either predefine query-associated videos without retrieval or convert videos into textual descriptions losing multimodal richness. To tackle these, we introduce VideoRAG, a framework that not only dynamically retrieves videos based on their relevance with queries and but also utilizes both visual and textual information. The operation of VideoRAG is powered by recent Large Video Language Models (LVLMs), which enable the direct processing of video content to represent it for retrieval and the seamless integration of retrieved videos jointly with queries for response generation. Also, inspired by that the context size of LVLMs may not be sufficient to process all frames in extremely long videos and not all frames are equally important, we introduce a video frame selection mechanism to extract the most informative subset of frames, along with a strategy to extract textual information from videos (as it can aid the understanding of video content) when their subtitles are not available. We experimentally validate the effectiveness of VideoRAG, showcasing that it is superior to relevant baselines.

VideoRAG: Retrieval-Augmented Generation over Video Corpus

Given a task in the form of a basic description and its training examples, prompt optimization is the problem of synthesizing the given information into a text prompt for a large language model. Humans solve this problem by also considering the different facets that define a task (e.g., counter-examples, explanations, analogies) and including them in the prompt. However, it is unclear whether existing algorithmic approaches, based on iteratively editing a given prompt or automatically selecting a few in-context examples, can cover the multiple facets required to solve a complex task. In this work, we view prompt optimization as that of learning multiple facets of a task from a set of training examples. We exploit structure in the prompt optimization problem and break down a prompt into loosely coupled semantic sections. The proposed algorithm, UniPrompt, (1) clusters the input space and uses clustered batches so that each batch likely corresponds to a different facet of the task, and (2) utilizes a feedback mechanism to propose adding, editing or deleting a section, which in turn is aggregated over a batch to capture generalizable facets. Empirical evaluation on multiple datasets and a real-world task shows that prompts generated using UniPrompt obtain higher accuracy than human-tuned prompts and those from state-of-the-art methods. In particular, our algorithm can generate long, complex prompts that existing methods are unable to generate.

Task Facet Learning: A Structured Approach To Prompt Optimization

To enhance reasoning capabilities, research has explored incorporating special-purpose tokens into the training process to strengthen the learning mechanism of transformer-based LLMs. Building on prior research that showed inserting dummy tokens consecutively just before reasoning steps can enhance effectiveness, we introduce a novel approach termed Dynamic Inserting tokens Training (DIT). Our method identifies positions within sequences where model confidence is lowest identified by token log likelihood, strategically inserting [PAUSE] tokens to bolster the model's predictive capabilities for subsequent tokens. Experimental results across diverse datasets and models, from 2.7B model to 8B model, demonstrate that DIT consistently outperforms traditional fine-tuning and previous token insertion methods. With this simple yet impactful method, we achieve accuracy gains of up to 4.7%p on GSM8K, 3.23%p on AQUA-RAT, and pass@1 improvements of up to 3.4%p on MBPP. Our work shows a model-based, dynamic approach rather than a heuristic one, thereby broadening the scope of research in reasoning.

Downloads

Next from ACL 2025

How Do LLMs Acquire New Knowledge? A Knowledge Circuits Perspective on Continual Pre-Training

Similar lecture

FACT-AUDIT: An Adaptive Multi-Agent Framework for Dynamic Fact-Checking Evaluation of Large Language Models