Datasets:
Formats:
json
Languages:
English
Size:
< 1K
Tags:
functional-metacognition
self-correction
reasoning
benchmark
error-recovery
declarative-procedural-gap
License:
Commit History
-
Duplicate from FINAL-Bench/Metacognitive
Co-authored-by: VIDRAFT_LAB <SeaWolf-AI@users.noreply.huggingface.co>18b0985