ViralLoop

Last week we released Claude Sonnet 4.5. As part of our alignment testing, we used a new tool to run automated audits for behaviors like sycophancy and deception. Now we’re open-sourcing the tool to run those audits.

Researchers give Petri a list of seed instructions targeting scenarios and behaviors they want to test. Petri then operates on each seed instruction in parallel. For each seed instruction, an auditor agent makes a plan and interacts with the target model in a tool use loop. At the end, a separate judge model scores each of the resulting transcripts across multiple fixed dimensions so researchers can quickly search and filter for the most interesting transcripts.

2.5K likes87 repliesPosted Oct 6, 2025

Remix this post

Your note

Private to you, and used when you remix this post.

Viral score

Based on absolute engagement — the platform doesn’t expose follower counts here, so large accounts score higher.

From likes, comments.

Why it worked

Drops you into the middle of a story, so the next line feels necessary.
Ends on a short line that is easy to quote and repeat.
The opening line stands alone, so it reads completely before the post is expanded.
Written from experience rather than as advice, which is harder to argue with.
Short enough to read in full without expanding, which removes a step before the point lands.
Short sentences keep the pace up, so the reader never has to re-read a line.

The mechanics

Hook: Opens mid-story
Structure: Hook → Landing
Style: Conversational paragraphs at a steady pace.
Tone: corporate
Written for: People building with AI

Analysis by ViralLoop — not written by the original author.

Post by Anthropic on X