Files
rlvrmathif-qwen2.5-1.5b/eval-results/ifbench/metrics.json

16 lines
435 B
JSON
Raw Normal View History

{
"ifbench": {
"pass@1": {
"num_prompts": 294,
"num_instructions": 335,
"average_score": 19.621027515483807,
"prompt_strict_accuracy": 17.006802721088434,
"instruction_strict_accuracy": 19.402985074626866,
"prompt_loose_accuracy": 19.387755102040817,
"instruction_loose_accuracy": 22.686567164179106,
"num_entries": 294,
"avg_tokens": 445,
"gen_seconds": 53
}
}
}