{ "label": "merged", "model": "/workspace/checkpoints/qwen36_a100_stage1_merged", "results": [ { "kind": "humaneval", "n": 164, "pass@1": 0.8171, "passed": 134 } ], "reference_pass@1": { "Qwen2.5-Coder-32B-Instruct": 0.92, "GPT-4o": 0.9, "DeepSeek-V3": 0.91, "Llama-3.1-70B-Instruct": 0.8, "Qwen2.5-7B-Instruct": 0.84, "CodeLlama-34B": 0.51 } }