Repository navigation
Add SciΠvot to Datasets and Benchmarks - #44
Merged
Merged
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 paper row, SciΠvot, to Datasets and Benchmarks (end of the main table, after AgentLeak).
Why it fits the scope: it checks whether an agent's decision at a key point of a research run is right. The 248 cases come from six months of real research, at points where senior researchers corrected an LLM agent (26 projects, six fields). Each case asks the model to pick the next step. Four of five frontier models score below random, and the models tend to keep the original LLM plan.
I added the row and then ran
tools/lintfix.py(it re-padded this one table) andtools/recount.py, which updated the derived figures in README.md andassets/social-card.html.recount.py --checkpasses. I did not re-rendersocial-card.png.