A supervised fine-tuning dataset for teaching language models to reason about
chess. It covers board perception, legal move generation, tactics, evaluation,
openings, endgames, and planning.
Total examples
15,100,000
Total size
18902.7 MB
Format
JSONL chat rows with messages
Eval companion
Chess-Nut-Engine/chess-sft-eval
Tier
Category
Tasks
Examples
Size
0
Unknown
1
300,000
556.2 MB
