China

The reliability of large language models (LLMs) is greatly compromised by their tendency to hallucinate, underscoring the need for precise identification of knowledge gaps within LLMs. Various methods for probing such gaps exist, ranging from calibration-based to prompting-based methods. To evaluate these probing methods, in this paper, we propose a new process based on using input variations and quantitative metrics. Through this, we expose two dimensions of inconsistency in knowledge gap probing. (1) **Intra-method inconsistency:** Minimal non-semantic perturbations in prompts lead to considerable variance in detected knowledge gaps within the same probing method; e.g., the simple variation of shuffling answer options can decrease agreement to around 40%. (2) **Cross-method inconsistency:** Probing methods contradict each other on whether a model knows the answer. Methods are highly inconsistent -- with decision consistency across methods being as low as 7% -- even though the model, dataset, and prompt are all the same. These findings challenge existing probing methods and highlight the urgent need for perturbation-robust probing frameworks.

EMNLP 2025

Do We Know What LLMs Don’t Know? A Study of Consistency in Knowledge Probing

poster

## 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."<br>

*Message from the General Chair, Dirk Hovy*

[**Link to Conference Handbook**](https://drive.google.com/file/d/1johU5QqVVYO4RfH7QcIORr7qrVBdzdwC/view?usp=sharing)





<br>

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

Large language models perform surprisingly well on many zero-shot classification tasks, but are difficult to fairly compare to supervised classifiers due to the lack of a modifiable decision boundary. In this work, we propose and evaluate a method that converts binary classification tasks into pairwise comparison tasks, obtaining relative rankings from LLMs. Repeated pairwise comparisons can be used to score instances using the Elo rating system (used in chess and other competitions), inducing a confidence ordering over instances in a dataset. We evaluate scheduling algorithms for their ability to minimize comparisons, and show that our proposed algorithm leads to improved classification performance, while also providing more information than traditional zero-shot classification.

Using tournaments to calculate AUROC for zero-shot classification with LLMs

Evaluating the factuality of long-form generations from Large Language Models (LLMs) remains challenging due to accuracy issues and costly human assessment. Prior evaluation pipelines attempt this by decomposing text into claims, searching for evidence, and verifying claims, but suffer from critical drawbacks: (1) inefficiency due to complex pipeline components unsuitable for long LLM outputs, and (2) ineffectiveness stemming from inaccurate claim sets and insufficient evidence collection of one-line SERP snippets. To address these limitations, we adapt the existing decompose-then-verify evaluation framework and propose **FaStFact**, a fast and strong evaluation pipeline that achieves the highest alignment with human evaluation and efficiency among existing baselines. FaStFact first employs chunk-level claim extraction integrated with confidence-based pre-verification, significantly reducing the cost of web searching and inference calling while ensuring reliability. For searching and verification, it gathers document-level evidence from crawled website pages for retrieval during verification, addressing the evidence insufficiency problem in previous pipelines. Extensive experiments based on an aggregated and manually annotated benchmark demonstrate the reliability of FaStFact in both efficiently and effectively evaluating the factuality of long-form LLM generations. We submit the paper with code and benchmark, and will make them publicly available to facilitate research.

FaStFact: Faster, Stronger Long-Form Factuality Evaluations in LLMs

There has been significant research on propagandistic content detection across different modalities and languages. However, most studies have primarily focused on detection, with little attention given to explanations justifying the predicted label. This is largely due to the lack of resources that provide explanations alongside annotated labels. To address this issue, we propose a multilingual (i.e., Arabic and English) explanation-enhanced dataset, the first of its kind. Additionally, we introduce an explanation-enhanced LLM for both label detection and rationale-based explanation generation. Our findings indicate that the model performs comparably while also generating explanations. We will make the dataset and experimental resources publicly available for the research community (anonymous.com).

PropXplain: Can LLMs Enable Explainable Propaganda Detection?

Knowledge-based complex reasoning remains a significant challenge for large language models (LLMs) with in-context learning. To tackle this issue, previous studies focus on ensuring behavior fidelity, factuality, or reliability in generated reasoning processes that guide LLMs to produce solutions. However, these studies often neglect the simultaneous optimization on all these three aspects for each thought. The main challenges are the lack of comprehensive assessment mechanisms and the difficulty of efficient thought-level optimization. This paper introduces the Evolution of Thoughts (EoT) framework, which enhances the factuality, fidelity, and reliability of each thought in the reasoning process through a few LLM inferences. We propose a thought assessment method that is sensitive to knowledge and LLM behaviors, using three scorers to evaluate each thought by considering domain context, semantic alignment, and behavior impact. Additionally, we establish a self-reflective evolution mechanism to facilitate each reasoning process generation in a single-forward inference. Extensive experiments demonstrate that, for knowledge-based complex tasks, EoT improves the factuality and fidelity of reasoning processes by approximately 16.5% and 48.8%, respectively, while enhancing LLM reasoning capability by about 6.2%, outperforming advanced approaches.

EoT: Evolution of Thoughts for Complex Reasoning Tasks

Large Language Models (LLMs) have revolutionized natural language processing, but their varying capabilities and costs pose challenges in practical applications. LLM routing addresses this by dynamically selecting the most suitable LLM for each query/task. Previous approaches treat this as a supervised learning problem, assuming complete knowledge of optimal query-LLM pairings. However, real-world scenarios lack such comprehensive mappings and face evolving user queries. We thus propose to study LLM routing as a contextual bandit problem, enabling adaptive decision-making using bandit feedback without requiring exhaustive inference across all LLMs for all queries (in contrast to supervised routing). To address this problem, we develop a shared embedding space for queries and LLMs, where query and LLM embeddings are aligned to reflect their affinity. This space is initially learned from offline human preference data and refined through online bandit feedback. We instantiate this idea through Preference-prior Informed Linucb fOr adaptive rouTing (PILOT), a novel extension of LinUCB. To handle diverse user budgets for model routing, we introduce an online cost policy modeled as a multi-choice knapsack problem, ensuring resource-efficient routing.

Adaptive LLM Routing under Budget Constraints

Agents powered by large language models (LLMs) have demonstrated strong planning and decision-making capabilities in complex embodied environments. However, such agents often suffer from inefficiencies in multi-turn interactions, frequently trapped in repetitive loops or issuing ineffective commands, leading to redundant computational overhead. Instead of relying solely on learning from trajectories, we take a first step toward exploring the early-exit behavior for LLM-based agents. We propose two complementary approaches, 1. an **intrinsic** method that injects exit instructions during generation, and 2. an **extrinsic** method that verifies task completion to determine when to halt an agent’s trial. To evaluate early-exit mechanisms, we introduce two metrics: one measures the reduction of **redundant steps** as a positive effect, and the other evaluates **progress degradation** as a negative effect. Experiments with 4 different LLMs across 5 embodied environments show significant efficiency improvements, with only minor drops in agent performance. We also validate a practical strategy where a stronger agent assists after an early-exit agent, achieving better performance with the same total steps. We will release our code to support further research.

Runaway is Ashamed, But Helpful: On the Early-Exit Behavior of Large Language Model-based Agents in Embodied Environments

Retrieval-augmented generation (RAG) enhances large language models (LLMs) with external knowledge but incurs significant inference costs due to lengthy retrieved contexts. While context compression mitigates this issue, existing methods apply fixed compression rates—over-compressing simple queries or under-compressing complex ones. We propose Adaptive Context Compression for RAG (ACC-RAG), a framework that dynamically adjusts compression rates based on input complexity, optimizing inference efficiency without sacrificing accuracy. ACC-RAG combines a hierarchical compressor (for multi-granular embeddings) with a context selector to retain minimal sufficient information, akin to human skimming. Evaluated on Wikipedia and five QA datasets, ACC-RAG outperforms fixed-rate methods and matches/unlocks >4× faster inference versus standard RAG while maintaining or improving accuracy.

Enhancing RAG Efficiency with Adaptive Context Compression

Scaling test-time computation—generating and analyzing multiple or sequential outputs for a single input—has become a promising strategy for improving the reliability and quality of large language models (LLMs), as evidenced by advances in uncertainty quantification and multi-step reasoning. A key shared component is semantic clustering, which groups outputs that differ in form but convey the same meaning. Semantic clustering enables estimation of the distribution over the semantics of outputs and helps avoid redundant exploration of reasoning paths. However, existing approaches typically rely on external models, which introduce substantial computational overhead and often fail to capture context-aware semantics. We propose Latent Semantic Clustering (LSC), a lightweight and context-sensitive method that leverages the generator LLM’s internal hidden states for clustering, eliminating the need for external models. Our extensive experiment across various LLMs and datasets shows that LSC significantly improves the computational efficiency of test-time scaling while maintaining or exceeding the performance of existing methods.

Efficient Latent Semantic Clustering for Scaling Test-Time Computation of LLMs

Web banner advertisements, which are placed on websites to guide users to a targeted landing page (LP), are still often selected manually because human preferences are important in selecting which ads to deliver. To automate this process, we propose a new benchmark, BannerBench, to evaluate the human preference-driven banner selection process using vision-language models (VLMs). This benchmark assesses the degree of alignment with human preferences in two tasks: a ranking task and a best-choice task, both using sets of five images derived from a single LP. Our experiments show that VLMs are moderately correlated with human preferences on the ranking task. In the best-choice task, most VLMs perform close to chance level across various prompting strategies. These findings suggest that although VLMs have a basic understanding of human preferences, most of them struggle to pinpoint a single suitable option from many candidates.

BannerBench: Benchmarking Vision Language Models for Multi-Ad Selection with Human Preferences

The exponential growth of scientific publications has overwhelmed reviewers and researchers, with top conferences receiving thousands of submissions annually. Reviewers must assess feasibility, novelty, and impact under tight deadlines, often lacking tools to identify relevant prior work. Early-career researchers face similar challenges, with limited support to navigate fast-evolving fields. Existing LLM-based systems struggle with static retrieval, surface-level features, and lack multi-hop reasoning, leading to shallow or hallucinated assessments. Scientific evaluation requires a deep, relational understanding, which current retrieval-augmented generation (RAG) methods fail to achieve. We introduce SciCompanion, a graph-grounded reasoning framework for structured scientific evaluation. Given a paper or abstract-like input, SciCompanion builds a dynamic knowledge graph from recent publications, domain-specific databases, and curated metadata. It employs multi-hop reasoning to iteratively construct contextual graphs and generate structured critiques, enabling deeper exploration of scientific literature. Unlike sentiment-biased LLM evaluations, SciCompanion directly optimizes retrieval and graph refinement using Group Relative Policy Optimization (GRPO), producing reviews aligned with expert judgments. Experiments on ICLR and ACL datasets show that SciCompanion reduces evaluation error by over 30% compared to prompting-only baselines and allows smaller models to outperform larger ones. Evaluations across three datasets, using metrics for retrieval accuracy, semantic overlap, and multi-hop sensitivity, along with a case study, demonstrate SciCompanion's robustness and versatility.

Downloads

Next from EMNLP 2025

Using tournaments to calculate AUROC for zero-shot classification with LLMs

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES