8 citations · 16 across the 28 of their papers we have counts for
32 papers
UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal Representations
Dvir Samuel, Guy Bar-Shalom, Fabrizio Frasca +4
Large Vision-Language Models (LVLMs) achieve impressive visual reasoning and dialogue capabilities, yet frequently hallucinate content unsupported by the visual input. Effective mi…
Paved with True Intents: Intent-Aware Training Improves LLM Safety Classification Across Training Regimes
Jeremias Ferrao, Niclas Müller-Hof, Iustin Sîrbu +2
We argue that safety classifiers should model user intent as an explicit signal between the prompt and the final label. To study this, we introduce AIMS, a human-annotated dataset…
Debiasing Without Protected Attributes: Latent Concept Erasure from Textual Profiles
Shun Shao, Zheng Zhao, Anna Korhonen +2
Most fairness research in NLP assumes direct access to protected attributes such as gender, race, or nationality. In practice, however, such information is often unavailable due to…
When Knowledge Is Not Free: Cost-Aware Evidence Selection in Retrieval-Augmented Generation
Mingyan Wu, Han Yang, Omer Ben-Porat +1
Retrieval-Augmented Generation (RAG) typically assumes that external knowledge is free, but many high-quality sources are paywalled, licensed, restricted, or otherwise costly to ac…
Self-Improving World Modelling with Latent Actions
Yifu Qiu, Zheng Zhao, Waylon Li +4
Internal modelling of the world -- predicting transitions between previous states and next states under actions -- is essential to reasoning and planning for LLMs and V…
From Competition to Collaboration: Designing Sustainable Mechanisms Between LLMs and Online Forums
Niv Fono, Yftah Ziser, Omer Ben-Porat
While Generative AI (GenAI) systems draw users away from (Q&A) forums, they also depend on the very data those forums produce to improve their performance. Addressing this paradox,…
Lost in Space? Vision-Language Models Struggle with Relative Camera Pose Estimation
Ken Deng, Yifu Qiu, Yoni Kasten +2
We study whether vision-language models (VLMs) can solve relative camera pose estimation (RCPE) from image pairs, a direct test of multi-view spatial reasoning. We cast RCPE as a d…
Beyond Token Probes: Hallucination Detection via Activation Tensors with ACT-ViT
Guy Bar-Shalom, Fabrizio Frasca, Yaniv Galron +2
Detecting hallucinations in Large Language Model-generated text is crucial for their safe deployment. While probing classifiers show promise, they operate on isolated layer-token p…
Neural Message-Passing on Attention Graphs for Hallucination Detection
Fabrizio Frasca, Guy Bar-Shalom, Yftah Ziser +1
Large Language Models (LLMs) often generate incorrect or unsupported content, known as hallucinations. Existing detection methods rely on heuristics or simple models over isolated…
Knowing Before Saying: LLM Representations Encode Information About Chain-of-Thought Success Before Completion
Anum Afzal, Florian Matthes, Gal Chechik +1
We investigate whether the success of a zero-shot Chain-of-Thought (CoT) process can be predicted before completion. We discover that a probing classifier, based on LLM representat…
Iterative Multilingual Spectral Attribute Erasure
Shun Shao, Yftah Ziser, Zheng Zhao +3
Multilingual representations embed words with similar meanings to share a common semantic space across languages, creating opportunities to transfer debiasing effects between langu…
Aligning What LLMs Do and Say: Towards Self-Consistent Explanations
Sahar Admoni, Ofra Amir, Assaf Hallak +1
Large language models (LLMs) seem to offer an easy path to interpretability: just ask them to explain their answers. Yet the features driving an answer often differ from those emph…
Can VLMs Predict Future States? Bootstrapping World Models from Inverse Dynamics
Yifu Qiu, Yftah Ziser, Anna Korhonen +2
Can unified vision-language models (VLMs) perform forward dynamics prediction (FDP), i.e., predicting the future state (in image form) given the previous observation and an action…
SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs
Shaona Ghosh, Amrita Bhattacharjee, Yftah Ziser +1
Fine-tuning large language models (LLMs) to adapt to evolving safety policies is costly and impractical. Mechanistic interpretability enables inference-time control through latent…
Policy Optimized Text-to-Image Pipeline Design
Uri Gadot, Rinon Gal, Yftah Ziser +2
Text-to-image generation has evolved beyond single monolithic models to complex multi-component pipelines. These combine fine-tuned generators, adapters, upscaling blocks and even…
Beyond Next Token Probabilities: Learnable, Fast Detection of Hallucinations and Data Contamination on LLM Output Distributions
Guy Bar-Shalom, Fabrizio Frasca, Derek Lim +5
The automated detection of hallucinations and training data contamination is pivotal to the safe deployment of Large Language Models (LLMs). These tasks are particularly challengin…
From Actions to Words: Towards Abstractive-Textual Policy Summarization in RL
Sahar Admoni, Assaf Hallak, Yftah Ziser +2
Explaining reinforcement learning agents is challenging because policies emerge from complex reward structures and neural representations that are difficult for humans to interpret…
TSPRank: Bridging Pairwise and Listwise Methods with a Bilinear Travelling Salesman Model
Weixian Waylon Li, Yftah Ziser, Yifei Xie +2
Traditional Learning-To-Rank (LETOR) approaches, including pairwise methods like RankNet and LambdaMART, often fall short by solely focusing on pairwise comparisons, leading to sub…
Layer by Layer: Uncovering Where Multi-Task Learning Happens in Instruction-Tuned Large Language Models
Zheng Zhao, Yftah Ziser, Shay B. Cohen
Fine-tuning pre-trained large language models (LLMs) on a diverse array of tasks has become a common approach for building models that can solve various natural language processing…
Spectral Editing of Activations for Large Language Model Alignment
Yifu Qiu, Zheng Zhao, Yftah Ziser +3
Large language models (LLMs) often exhibit undesirable behaviours, such as generating untruthful or biased content. Editing their internal representations has been shown to be effe…
Are Large Language Models Temporally Grounded?
Yifu Qiu, Zheng Zhao, Yftah Ziser +3
Are Large language models (LLMs) temporally grounded? Since LLMs cannot perceive and interact with the environment, it is impossible to answer this question directly. Instead, we p…
A Joint Matrix Factorization Analysis of Multilingual Representations
Zheng Zhao, Yftah Ziser, Bonnie Webber +1
We present an analysis tool based on joint matrix factorization for comparing latent representations of multilingual and monolingual models. An alternative to probing, this tool al…
Detecting and Mitigating Hallucinations in Multilingual Summarisation
Yifu Qiu, Yftah Ziser, Anna Korhonen +2
Hallucinations pose a significant challenge to the reliability of neural models for abstractive summarisation. While automatically generated summaries may be fluent, they often lac…
BERT is not The Count: Learning to Match Mathematical Statements with Proofs
Weixian Waylon Li, Yftah Ziser, Maximin Coavoux +1
We introduce a task consisting in matching a proof to a given mathematical statement. The task fits well within current research on Mathematical Information Retrieval and, more gen…