45 citations · 302 across the 119 of their papers we have counts for
33 papers · 1 filter
Don't Throw Away Your Beams: Improving Consistency-based Uncertainties in LLMs via Beam Search
Ekaterina Fadeeva, Maiya Goloburda, Aleksandr Rubashevskii +5
Consistency-based methods have emerged as an effective approach to uncertainty quantification (UQ) in large language models. These methods typically rely on several generations obt…
ReProbe: Efficient Test-Time Scaling of Multi-Step Reasoning by Probing Internal States of Large Language Models
Jingwei Ni, Ekaterina Fadeeva, Tianyi Wu +8
LLMs can solve complex tasks by generating long, multi-step reasoning chains. Test-time scaling (TTS) can further improve performance by sampling multiple variants of intermediate…
On the Emergence of Induction Heads for In-Context Learning
Tiberiu Musat, Tiago Pimentel, Lorenzo Noci +3
Transformers have become the dominant architecture for natural language processing. Part of their success is owed to a remarkable capability known as in-context learning (ICL): the…
Sample Smart, Not Hard: Correctness-First Decoding for Better Reasoning in LLMs
Xueyan Li, Guinan Su, Mrinmaya Sachan +1
Large Language Models (LLMs) are increasingly applied to complex tasks that require extended reasoning. In such settings, models often benefit from diverse chains-of-thought to arr…
Are Language Models Efficient Reasoners? A Perspective from Logic Programming
Andreas Opedal, Yanick Zengaffinen, Haruki Shirakami +5
Modern language models (LMs) exhibit strong deductive reasoning capabilities, yet standard evaluations emphasize correctness while overlooking a key aspect of reasoning: efficiency…
Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
Ziheng Chi, Yifan Hou, Chenxi Pang +3
Diagrams convey symbolic information in a visual format rather than a linear stream of words, making them especially challenging for AI models to process. While recent evaluations…
Can Vision-Language Models Solve Visual Math Equations?
Monjoy Narayan Choudhury, Junling Wang, Yifan Hou +1
Despite strong performance in visual understanding and language-based reasoning, Vision-Language Models (VLMs) struggle with tasks requiring integrated perception and symbolic comp…
Compose and Fuse: Revisiting the Foundational Bottlenecks in Multimodal Reasoning
Yucheng Wang, Yifan Hou, Aydin Javadov +2
Multimodal large language models (MLLMs) promise enhanced reasoning by integrating diverse inputs such as text, vision, and audio. Yet cross-modal reasoning remains underexplored,…
Test of Time: Rethinking Temporal Signal of Benchmark Contamination
Terry Jingchen Zhang, Gopal Dev, Ning Wang +8
Post-cutoff performance decay of LLMs has been widely interpreted as a temporal signal for benchmark contamination, where public information released before the training cutoff may…
COMET-poly: Machine Translation Metric Grounded in Other Candidates
Maike Züfle, Vilém Zouhar, Tu Anh Dinh +3
Automated metrics for machine translation attempt to replicate human judgment. Unlike humans, who often assess a translation in the context of multiple alternatives, these metrics…
Lean Meets Theoretical Computer Science: Scalable Synthesis of Theorem Proving Challenges in Formal-Informal Pairs
Terry Jingchen Zhang, Wenyuan Jiang, Rongchuan Liu +6
Formal theorem proving (FTP) has emerged as a critical foundation for evaluating the reasoning capabilities of large language models, enabling automated verification of mathematica…
Corrupted by Reasoning: Reasoning Language Models Become Free-Riders in Public Goods Games
David Guzman Piedrahita, Yongjin Yang, Mrinmaya Sachan +3
As large language models (LLMs) are increasingly deployed as autonomous agents, understanding their cooperation and social mechanisms is becoming increasingly important. In particu…
Probing for Arithmetic Errors in Language Models
Yucheng Sun, Alessandro Stolfo, Mrinmaya Sachan
We investigate whether internal activations in language models can be used to detect arithmetic errors. Starting with a controlled setting of 3-digit addition, we show that simple…
Personalized Exercise Recommendation with Semantically-Grounded Knowledge Tracing
Yilmazcan Ozyurt, Tunaberk Almaci, Stefan Feuerriegel +1
We introduce ExRec, a general framework for personalized exercise recommendation with semantically-grounded knowledge tracing. Our method builds on the observation that existing ex…
Co-DETECT: Collaborative Discovery of Edge Cases in Text Classification
Chenfei Xiong, Jingwei Ni, Yu Fan +10
We introduce Co-DETECT (Collaborative Discovery of Edge cases in TExt ClassificaTion), a novel mixed-initiative annotation framework that integrates human expertise with automatic…
The Medium Is Not the Message: Deconfounding Document Embeddings via Linear Concept Erasure
Yu Fan, Yang Tian, Shauli Ravfogel +3
Embedding-based similarity metrics between text sequences can be influenced not just by the content dimensions we most care about, but can also be biased by spurious attributes lik…
Educators' Perceptions of Large Language Models as Tutors: Comparing Human and AI Tutors in a Blind Text-only Setting
Sankalan Pal Chowdhury, Terry Jingchen Zhang, Donya Rooein +3
The rapid development of Large Language Models (LLMs) opens up the possibility of using them as personal tutors. This has led to the development of several intelligent tutoring sys…
Generating Pedagogically Meaningful Visuals for Math Word Problems: A New Benchmark and Analysis of Text-to-Image Models
Junling Wang, Anna Rutkiewicz, April Yi Wang +1
Visuals are valuable tools for teaching math word problems (MWPs), helping young learners interpret textual descriptions into mathematical expressions before solving them. However,…
Can Reasoning Help Large Language Models Capture Human Annotator Disagreement?
Jingwei Ni, Yu Fan, Vilém Zouhar +6
Variation in human annotation (i.e., disagreements) is common in NLP, often reflecting important information like task subjectivity and sample ambiguity. Modeling this variation is…
Dense SAE Latents Are Features, Not Bugs
Xiaoqing Sun, Alessandro Stolfo, Joshua Engels +4
Sparse autoencoders (SAEs) are designed to extract interpretable features from language models by enforcing a sparsity constraint. Ideally, training an SAE would yield latents that…
Improving Large Language Model Safety with Contrastive Representation Learning
Samuel Simko, Mrinmaya Sachan, Bernhard Schölkopf +1
Large Language Models (LLMs) are powerful tools with profound societal impacts, yet their ability to generate responses to diverse and uncontrolled inputs leaves them vulnerable to…
A Head to Predict and a Head to Question: Pre-trained Uncertainty Quantification Heads for Hallucination Detection in LLM Outputs
Artem Shelmanov, Ekaterina Fadeeva, Akim Tsvigun +9
Large Language Models (LLMs) have the tendency to hallucinate, i.e., to sporadically generate false or fabricated information. This presents a major challenge, as hallucinations of…
Faithfulness-Aware Uncertainty Quantification for Fact-Checking the Output of Retrieval Augmented Generation
Ekaterina Fadeeva, Aleksandr Rubashevskii, Dzianis Piatrashyn +7
Large Language Models (LLMs) enhanced with retrieval, an approach known as Retrieval-Augmented Generation (RAG), have achieved strong performance in open-domain question answering.…
Efficient Hallucination Detection for LLMs Using Uncertainty-Aware Attention Heads
Artem Vazhentsev, Lyudmila Rvanova, Gleb Kuzmin +8
While large language models (LLMs) have become highly capable, they remain prone to factual inaccuracies, commonly referred to as "hallucinations." Uncertainty quantification (UQ)…