{ "checkpoint": "/Users/paras/projects/mempool/research/models/20260628-qwen-small-logits-orchestrator-smoke/qwen_logits_heads.pt", "mean_worker_loss": 1.4604794980788773, "mean_workflow_loss": 0.41937757511605567, "prediction_sample": [ { "predicted_worker_id": "ollama-cloud-kimi-k2.7-code", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "verify_then_fallback" }, { "predicted_worker_id": "ollama-cloud-qwen3-coder-480b", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-kimi-k2.7-code", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-qwen3-coder-480b", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-kimi-k2.7-code", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-deepseek-v4-pro", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-kimi-k2.7-code", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-glm-5.2", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "verify_then_fallback" }, { "predicted_worker_id": "ollama-cloud-deepseek-v4-pro", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-deepseek-v4-pro", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-deepseek-v4-pro", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-qwen3-coder-480b", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-kimi-k2.7-code", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "direct" }, { "predicted_worker_id": "ollama-cloud-qwen3-coder-480b", "predicted_workflow": "direct", "target_worker_id": "ollama-cloud-qwen3-coder-480b", "target_workflow": "verify_then_fallback" } ], "record_count": 66, "rows": "/Users/paras/projects/mempool/research/datasets/20260628-qwen-small-logits-orchestrator-smoke-rows.jsonl", "schema_version": "mempool.qwen_logits_orchestrator_eval_report.v1", "worker_accuracy": 0.3181818181818182, "worker_ids": [ "ollama-cloud-deepseek-v4-pro", "ollama-cloud-glm-5.2", "ollama-cloud-kimi-k2.7-code", "ollama-cloud-qwen3-coder-480b" ], "workflow_accuracy": 0.8636363636363636, "workflow_labels": [ "direct", "verify_then_fallback" ] }