5 citations · 10 across the 12 of their papers we have counts for
13 papers
Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble
Jorio Cocola, Lev McKinney, Harry Mayne +2
Language models are trained to implement a helpful AI Assistant character (e.g., Claude). We explore how finetuning on synthetic stories affects this character. Does it change the…
Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values
Jan Betley, Johannes Treutlein, Jan Dubiński +5
People use language models for practical questions whose answers are difficult to verify. We show that models exhibit covert value leakage: the information they provide is influenc…
Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models
Dewi Gould, Francis Rhys Ward, Anders Cairns Woodruff +18
Many efforts to ensure frontier AI models are safe rely on monitoring their chain-of-thought (CoT) reasoning. If models become able to perform sufficiently complex reasoning intern…
Negation Neglect: When models fail to learn negations in training
Harry Mayne, Lev McKinney, Jan Dubiński +3
We introduce Negation Neglect, where finetuning LLMs on documents that flag a claim as false makes them believe the claim is true. For example, models are finetuned on documents th…
A Positive Case for Faithfulness: LLM Self-Explanations Help Predict Model Behavior
Harry Mayne, Justin Singh Kang, Dewi Gould +3
LLM self-explanations are often presented as a promising tool for AI oversight, yet their faithfulness to the model's true reasoning process is poorly understood. Existing faithful…
Measuring what Matters: Construct Validity in Large Language Model Benchmarks
Andrew M. Bean, Ryan Othniel Kearns, Angelika Romanou +39
Evaluating large language models (LLMs) is crucial for both assessing their capabilities and identifying safety or robustness issues prior to deployment. Reliably measuring abstrac…
LLMs Don't Know Their Own Decision Boundaries: The Unreliability of Self-Generated Counterfactual Explanations
Harry Mayne, Ryan Othniel Kearns, Yushi Yang +4
To collaborate effectively with humans, language models must be able to explain their decisions in natural language. We study a specific type of self-explanation: self-generated co…
LINGOLY-TOO: Disentangling Reasoning from Knowledge with Templatised Orthographic Obfuscation
Jude Khouja, Lingyi Yang, Karolina Korgul +6
Frontier language models demonstrate increasing ability at solving reasoning problems, but their performance is often inflated by circumventing reasoning and instead relying on the…
Can sparse autoencoders be used to decompose and interpret steering vectors?
Harry Mayne, Yushi Yang, Adam Mahdi
Steering vectors are a promising approach to control the behaviour of large language models. However, their underlying mechanisms remain poorly understood. While sparse autoencoder…
How Does DPO Reduce Toxicity? A Mechanistic Neuron-Level Analysis
Yushi Yang, Filip Sondej, Harry Mayne +2
Safety fine-tuning algorithms reduce harmful outputs in language models, yet their mechanisms remain under-explored. Direct Preference Optimization (DPO) is a popular choice of alg…
LINGOLY: A Benchmark of Olympiad-Level Linguistic Reasoning Puzzles in Low-Resource and Extinct Languages
Andrew M. Bean, Simi Hellsten, Harry Mayne +5
In this paper, we present the LingOly benchmark, a novel benchmark for advanced reasoning abilities in large language models. Using challenging Linguistic Olympiad puzzles, we eval…
Large language models can help boost food production, but be mindful of their risks
Djavan De Clercq, Elias Nehring, Harry Mayne +1
Coverage of ChatGPT-style large language models (LLMs) in the media has focused on their eye-catching achievements, including solving advanced mathematical problems and reaching ex…
Unsupervised Learning Approaches for Identifying ICU Patient Subgroups: Do Results Generalise?
Harry Mayne, Guy Parsons, Adam Mahdi
The use of unsupervised learning to identify patient subgroups has emerged as a potentially promising direction to improve the efficiency of Intensive Care Units (ICUs). By identif…