Circuits Updates - April 2025

We report a number of developing ideas on the Anthropic interpretability team, which might be of interest to researchers working actively in this space. Some of these are emerging strands of research where we expect to publish more on in the coming months. Others are minor points we wish to share, since we're unlikely to ever write a paper about them. We'd ask you to treat these results like those of a colleague sharing some thoughts or preliminary experiments for a few minutes at a lab meeting, rather than a mature paper. New Posts In “On the Biology of a Large Language Model”, we examined a jailbreak in which Haiku partially complies with a request to explain how to make a bomb, and compared it to a baseline prompt in which Haiku is directly asked for instructions for making a bomb, but refuses. One prompt we examined but did not include in the paper is an unsuccessful jailbreak which we discuss below. Surprisingly, by applying the same circuit tracing methodology to this prompt, we

Circuits Updates - April 2025 Transformer Circuits Thread Circuits Updates - April 2025 We report a number of developing ideas on the Anthropic interpretability team, which might be of interest to researchers working actively in this space. Some of these are emerging strands of research where we expect to publish more on in the coming months. Others are minor points we wish to share, since we're unlikely to ever write a paper about them. We'd ask you to treat these results like those of a colleague sharing some thoughts or preliminary experiments for a few minutes at a lab meeting, rather than

related reading

• Transformer Circuits Threadtransformer-circuits.pub
• On the Biology of a Large Language Modeltransformer-circuits.pub
• Circuits Updates - April 2024transformer-circuits.pub
• Circuits Updates - July 2024transformer-circuits.pub
• Circuits Updates - January 2024transformer-circuits.pub
• Circuit Tracing: Revealing Computational Graphs in Language Modelstransformer-circuits.pub
• Circuits Updates — May 2023transformer-circuits.pub
• Tracing the thoughts of a large language model \ Anthropicanthropic.com
• Circuits Updates - February 2024transformer-circuits.pub
• A Mathematical Framework for Transformer Circuitstransformer-circuits.pub
• Towards Automated Circuit Discovery for Mechanistic Interpretabilityarxiv.org
• An Ambitious Vision for Interpretability — AI Alignment Forumalignmentforum.org