RepoJournal
Safety Evals Safety Evals
24 wires and counting

$ follow Safety Evals

Keep up with Safety Evals in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-10-19
stories 1
RSS Archive Reviews

© 2026 RepoJournal Home Showcase How it works Privacy

$ the-wire · showcase

Inspect AI ships an extensions and evals refresh

By RepoJournal · Filed · About Safety Evals · Composed from the cited sources · methodology

J.J. Allaire
1 person shipped this
J.J. Allaire @jjallaire 1 cited

A single maintenance commit to UKGovernmentBEIS/inspect_ai, described only as "update extensions and evals", landed with no accompanying release notes to say what operators should change.

update extensions and evals UKGovernmentBEIS/inspect_ai

by J.J. Allaire

The commit, authored by J.J. Allaire, is titled "update extensions and evals" and carries changes to both the extensions and evals surfaces of the repository, though the commit summary gives no detail on what those changes do or whether existing eval definitions keep running unchanged. Pin your Inspect AI version if a run depends on a specific extension or eval behavior, and read the diff befor...

$ ls ai-safety-evals/reviews/ # the step-back reads

the week in review Inspect 0.3.282 and 0.3.283 both break configs
Full archive → Weekly reviews →

Quick answers

What shipped in Safety Evals on October 19, 2026?
A single maintenance commit to UKGovernmentBEIS/inspect_ai, described only as "update extensions and evals", landed with no accompanying release notes to say what operators should change. In total, 1 commits landed.
Who contributed to Safety Evals on October 19, 2026?
1 developer shipped this update, including J.J. Allaire.
What were the notable Safety Evals updates?
update extensions and evals.
♥

$ log Safety Evals --recent

Viewer plays terminal sessions, Kimi K3.1 Pro lands in 0.3.284 13 shipped inspect_evals 0.26.0 drops GAIA's browser tools; background subagent failures now fail samples 61 shipped Inspect logs stop mutating mid-run, Math.Calc regrades 23 shipped Inspect 0.3.283 breaks datasets that use a `checkpoint` field 38 shipped

$ log --all # elsewhere on the wire

Alignment & Post-training GSM8K scoring parses answers as exact decimals AI Red Teaming PyRIT closes a judge-spoofing hole, promptfoo accepts Langfuse v6

$ ls showcases/

AI Red Teaming Agent Security Interpretability Alignment & Post-training

$ init your-desk

RepoJournal can write one of these for any repo you care about. Sign in with GitHub, pick your repos and sources, and your desk gets a wire like this one every morning.