Summary We’re a clinically vulnerable family and we take infection control very seriously. My first-author paper was accepted as a spotlight (top 23 out of 801) at the ICML 2026 Mechanistic Interpretability Workshop. We thought long and hard and decided that me travelling from the UK to Seoul is worth...
TL;DR We apply mechinterp techniques on VPT, OpenAI's Minecraft agent. We also find a new case of goal misgeneralization - VPT kills a villager when we force one to stand under some tree leaves. Abstract > Understanding the mechanisms behind decisions taken by large foundation models in sequential decision making...
TL;DR Different misgeneralized goals can appear based on training random seed alone, including outliers on a scale of 1 in 500. Abstract > We explore colour versus shape goal misgeneralization originally demonstrated > by Di Langosco et al. (2022) in the Procgen Maze environment, where, given > an ambiguous choice,...