AIConfPaper
← Search

Clement Neo

7 accepted papers

2025

Noise Injection Reveals Hidden Capabilities of Sandbagging Language Models

Cameron Tice, Philipp Alexander Kreer, Nathan Helm-Burger, Prithviraj Singh Shahani, Fedor Ryzhenkov, Fabien Roger +4
NeurIPS 2025poster

Capability evaluations play a crucial role in assessing and regulating frontier AI systems. The effectiveness of these evaluations faces a significant challenge: strategic underperformance, or ``sandbagging'', where models deliberately underperform during evaluation. Sandbagging can manifest either…

Cited by 0Source ↗code ↗Scholar ↗
2025

TinySQL: A Progressive Text-to-SQL Dataset for Mechanistic Interpretability Research

Abir Harrasse, Philip Quirke, Clement Neo, Dhruv Nathawani, Luke Marks, Amir Abdullah
EMNLP 2025

Mechanistic interpretability research faces a gap between analyzing simple circuits in toy tasks and discovering features in large models. To bridge this gap, we propose text-to-SQL generation as an ideal task to study, as it combines the formal structure of toy tasks with real-world complexity. We

Cited by 0Source ↗PDF ↗Scholar ↗
2025

Towards Interpreting Visual Information Processing in Vision-Language Models

Clement Neo, Luke Ong, Philip Torr, Mor Geva, David Krueger, Fazl Barez
ICLR 2025poster

Vision-Language Models (VLMs) are powerful tools for processing and understanding text and images. We study the processing of visual tokens in the language model component of LLaVA, a prominent VLM. Our approach focuses on analyzing the localization of object information, the evolution of visual tok…

Cited by 12Source ↗PDF ↗code ↗Scholar ↗
2025

Turning Up the Heat: Min-p Sampling for Creative and Coherent LLM Outputs

Nguyen Nhat Minh, Andrew Baker, Clement Neo, Allen G Roush, Andreas Kirsch, Ravid Shwartz-Ziv
ICLR 2025oral

Large Language Models (LLMs) generate text by sampling the next token from a probability distribution over the vocabulary at each decoding step. Popular sampling methods like top-p (nucleus sampling) often struggle to balance quality and diversity, especially at higher temperatures which lead to inc…

Cited by 8Source ↗PDF ↗Scholar ↗
2025

Understanding Refusal in Language Models with Sparse Autoencoders

Wei Jie Yeo, Nirmalendu Prakash, Clement Neo, Ranjan Satapathy, Roy Ka-Wei Lee, Erik Cambria
EMNLP 2025

Refusal is a key safety behavior in aligned language models, yet the internal mechanisms driving refusals remain opaque. In this work, we conduct a mechanistic study of refusal in instruction-tuned LLMs using sparse autoencoders to identify latent features that causally mediate refusal behaviors. We

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2024

Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions

Clement Neo, Shay B Cohen, Fazl Barez
EMNLP 2024main

Understanding the inner workings of large language models (LLMs) is crucial for advancing their theoretical foundations and real-world applications. While the attention mechanism and multi-layer perceptrons (MLPs) have been studied independently, their interactions remain largely unexplored. This st…

Cited by 1Source ↗PDF ↗Scholar ↗
2024

Interpreting Learned Feedback Patterns in Large Language Models

Luke Marks, Amir Abdullah, Clement Neo, Rauno Arike, David Krueger, Philip Torr +1
NeurIPS 2024poster

Reinforcement learning from human feedback (RLHF) is widely used to train large language models (LLMs). However, it is unclear whether LLMs accurately learn the underlying preferences in human feedback data. We coin the term **Learned Feedback Pattern** (LFP) for patterns in an LLM's activations lea…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗