Add: Breaking Agent Backbones paper (ICLR 2026) - #41
Open
jb-lakera wants to merge 1 commit into
Open
Conversation
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Our work recently has been accepted to ICLR and I believe it could be of interest!
The PR adds the paper Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents to the Security & Discussion → Papers section.
What the paper does: Introduces "threat snapshots" — a framework identifying specific execution points where LLM vulnerabilities manifest and propagate to agent-level risks. Presents the b³ benchmark, built from 194,331 crowdsourced adversarial attacks, evaluating 34 LLMs. Key finding: reasoning capability (not model size) correlates with security. Benchmark, dataset, and evaluation code are publicly released.
Thanks for reviewing!