Adam Jermyn
@AdamSJermyn
AI Interpretability & Safety @AnthropicAI. Previously at @FlatironInst @FlatironCCA, @KITP_UCSB, PhD @Cambridge_Uni, BS @Caltech.
A collection of small updates from the Anthropic Interpretability team: transformer-circuits.pub/2025/july-upda…
Who knew you could win gold in the International Math Olympiad without truly reasoning?
we've got some really cool new work looking at @AnthropicAI's circuit tracing work and comparing it to a circuit that's already been studied. some really interesting findings in here
New research update! We replicated @AnthropicAI's circuit tracing methods to test if they can recover a known, simple transformer mechanism.
Wild new examples of OpenAI shadiness. Great scoop
SCOOP: I obtained a previously unreported letter from OpenAI to California's Attorney General that includes surprising admissions about the company's restructuring plans – and shows how OpenAI is attacking critics who question its attempts to diminish its nonprofit governance.
A collection of small updates from the Anthropic Interpretability team (transformer-circuits.pub/2025/april-upd…). [Note this is different from the Attention update... it's been a busy time!]
An update on Attention from the Anthropic Interpretability team (transformer-circuits.pub/2025/attention…). This one is close to home since understanding attention is a lot of what I've worked on the last ... 18 months.