Add Provael under Data, Simulation & Benchmarks - #1
Open
sattyamjjain wants to merge 1 commit into
Open
sattyamjjain wants to merge 1 commit into
sattyamjjain wants to merge 1 commit into
Conversation
The section covers capability benchmarks; this is the adversarial counterpart. Software rather than a paper, stated in the entry since every other row is an arXiv link. Format follows CONTRIBUTING: - [Title](URL) - note (year).
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds one entry to 🧪 Data, Simulation & Benchmarks, in the section's format:
- [Title](URL) — note (year).Disclosure: I am the author.
The section covers benchmarks that measure capability — LIBERO, SIMPLER, CALVIN, RoboCasa, VLABench. This is the adversarial counterpart: it perturbs the instruction and observation a policy receives inside LIBERO or Meta-World and asks whether the end-effector left a keep-out region, reporting the rate with a task-clustered 95% CI against a matched benign control. The list currently has no safety or security entry at all, which is why I thought it was worth proposing rather than a stretch.
Software, not a paper. Every other row links an arXiv id; this has none, and the entry says so rather than implying a publication. If a paperless row does not fit the section, closing this is the right call.
Limits: simulation only, one policy measured on one suite, low single-digit stars, two months old. Two of its three measured attack families scored 0% against the real model and that is published at the same prominence as the positive result.
Link verified resolving.