Interactive tool

R. Luger; Harish Kamath; Doug Finkbeiner; Purvi Goel; Adam Jermyn; Sam Zimmerman; Joshua Batson; Tom Conerly · Anthropic

An interactive tool for interrogating attention heads, released as the research output in its own right. Its central finding is methodological: a head's behaviour on the full data distribution rarely matches what a narrow task suggests.

Reframes the tool as the contribution, and supplies the concrete examples the field needs to attack attention decomposition.

High · Read the full article, 2026-08-06

Aeree Cho; Grace C. Kim; Alexander Karpekov; Alec Helbling; Zijie J. Wang; Seongmin Lee; Benjamin Hoover; Duen Horng (Polo) Chau · Georgia Institute of Technology (Polo Club of Data Science)

An interactive tool that runs a live GPT-2 instance in your browser and visualises every stage of the forward pass — embeddings, attention, MLP, output probabilities — updating as you type your own text.

The most accessible route into transformer internals that exists: no installation, no GPU, no prior knowledge. That is a different kind of contribution from any research result in this ledger, and arguably a wider-reaching one.

High · Read the arXiv abstract page in full, 2026-08-06. The live tool itself was not inspected — it is client-rendered and the fetch returned an empty page shell.

Catherine Yeh; Yida Chen; Aoyu Wu; Cynthia Chen; Fernanda Viégas; Martin Wattenberg · Harvard University

Visualises a joint embedding of the query and key vectors a transformer uses to compute attention, which makes it possible to see global patterns across many input sequences rather than one prompt at a time.

Shifted attention visualisation from single-example to distribution-level — the exact move Anthropic's HeadVis names as its closest prior work three years later.

High · Read the arXiv abstract page in full, 2026-08-06. The live demo at attentionviz.com was not inspected — client-rendered, fetch returned an empty page shell.