11 citations · 23 across the 12 of their papers we have counts for
11 papers · 1 filter
Debate Training Reduces Reward Hacking in RLAIF
Zachary Kenton, Lili Janzer, Rory Greig +8
We demonstrate that RL finetuning an LLM using debate, a two-player adversarial game between a generator and a critic adjudicated by a weaker LLM judge, reduces reward hacking comp…
How Transparent is DiffusionGemma?
Joshua Engels, Callum McDougall, Bilal Chughtai +11
LLM reasoning transparency is a critical affordance for understanding model decisions, mitigating misuse and misalignment, and debugging surprising model behaviors. However, Diffus…
Building Production-Ready Probes For Gemini
János Kramár, Joshua Engels, Zheng Wang +4
Frontier language model capabilities are improving rapidly. We thus need stronger mitigations against bad actors misusing increasingly powerful systems. Prior work has shown that a…
Evaluating Sparse Autoencoders for Monosemantic Representation
Moghis Fereidouni, Muhammad Umair Haider, Peizhong Ju +1
A key barrier to interpreting large language models is polysemanticity, where neurons activate for multiple unrelated concepts. Sparse autoencoders (SAEs) have been proposed to mit…
MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking
Sebastian Farquhar, Vikrant Varma, David Lindner +4
Future advanced AI systems may learn sophisticated strategies through reinforcement learning (RL) that humans cannot understand well enough to safely evaluate. We propose a trainin…
Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
Tom Lieberum, Senthooran Rajamanoharan, Arthur Conmy +7
Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network's latent representations into seemingly interpretable features. Despit…
On scalable oversight with weak LLMs judging strong LLMs
Zachary Kenton, Noah Y. Siegel, János Kramár +8
Scalable oversight protocols aim to enable humans to accurately supervise superhuman AI. In this paper we study debate, where two AI's compete to convince a judge; consultancy, whe…
Improving Dictionary Learning with Gated Sparse Autoencoders
Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith +5
Recent work has found that sparse autoencoders (SAEs) are an effective technique for unsupervised discovery of interpretable features in language models' (LMs) activations, by find…
Evaluating Frontier Models for Dangerous Capabilities
Mary Phuong, Matthew Aitchison, Elliot Catt +24
To understand the risks posed by a new AI system, we must understand what it can and cannot do. Building on prior work, we introduce a programme of new "dangerous capability" evalu…
AtP*: An efficient and scalable method for localizing LLM behaviour to components
János Kramár, Tom Lieberum, Rohin Shah +1
Activation Patching is a method of directly computing causal attributions of behavior to model components. However, applying it exhaustively requires a sweep with cost scaling line…
Challenges with unsupervised LLM knowledge discovery
Sebastian Farquhar, Vikrant Varma, Zachary Kenton +3
We show that existing unsupervised methods on large language model (LLM) activations do not discover knowledge -- instead they seem to discover whatever feature of the activations…