Is this your dataset? Claim it with the Hugging Face account that owns it.

maveryn/
trace-eval-runs

Name: trace-eval-runs
Creator: maveryn

Views

No views yet

TRACE Evaluation Runs

This repository contains the canonical response, extraction, and scoring artifacts for TRACE validation and trace_eval_v1, the 24-benchmark external transfer evaluation used in the paper. Each run records model identities, decoding settings, content hashes, and aggregate scores. Paper · Project page · GitHub · Collection · Evaluation code · External-transfer results · TRACE validation results · Benchmark provenance

TRACE validation

The… See the full description on the dataset page: https://huggingface.co/datasets/maveryn/trace-eval-runs.