{
  "model_details": {
    "provider": "requesty",
    "model_requested": "claude-sonnet-4-6",
    "model_for_requests": "claude-sonnet-4-6",
    "api_base_url": "https://router.requesty.ai/v1",
    "chat_completions_endpoint": "https://router.requesty.ai/v1/chat/completions"
  },
  "cache_padding": {
    "enabled": true,
    "target_shared_prefix_tokens": 1200,
    "calibration_shared_prefix_tokens": 936,
    "target_prompt_tokens": 1200,
    "calibration_prompt_tokens": 936,
    "calibration_example_id": "119",
    "applied_padding_tokens_estimate": 106,
    "examples_with_padding_applied": 10
  },
  "request_control_summary": {
    "configured": {
      "requesty_auto_cache": true
    },
    "attempts_total": 1088,
    "attempts_with_control_telemetry": 1088,
    "per_control": {
      "reasoning_effort": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "thinking_level": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "effort": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "verbosity": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "prompt_cache_key": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "gemini_cached_content": {
        "configured_value": null,
        "requested_attempts": 0,
        "sent_attempts": 0,
        "accepted_attempts": 0,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": null,
        "rejected_reasons": {},
        "rejected_example_ids": []
      },
      "requesty_auto_cache": {
        "configured_value": true,
        "requested_attempts": 1088,
        "sent_attempts": 1088,
        "accepted_attempts": 1088,
        "rejected_attempts": 0,
        "missing_from_final_request_attempts": 0,
        "acceptance_rate": 1.0,
        "rejected_reasons": {},
        "rejected_example_ids": []
      }
    }
  },
  "usage_metadata_summary": {
    "attempts_total": 1088,
    "attempts_with_usage_metadata": 1088,
    "attempts_with_cached_token_signals": 1088,
    "cached_tokens_total_estimate": 0,
    "cache_read_tokens_total": 0,
    "cache_write_tokens_total": 0,
    "cache_token_fields_totals": {
      "usage.prompt_tokens_details.cached_tokens": 0
    },
    "attempts_with_gemini_cached_content_token_signals": 0,
    "gemini_cached_content_token_count_total": 0,
    "gemini_cached_content_token_fields_totals": {}
  },
  "token_usage_totals": {
    "attempts_total": 1088,
    "attempts_with_token_usage": 1000,
    "attempts_with_output_tokens": 1000,
    "attempts_with_cached_input_tokens": 0,
    "attempts_with_thinking_tokens": 0,
    "input_tokens_total": 1638142,
    "cached_input_tokens_total": 0,
    "non_cached_input_tokens_total": 1638142,
    "output_tokens_total": 182453,
    "thinking_tokens_total": 0,
    "output_tokens_definition": "total_tokens - prompt_tokens (or completion_tokens + thinking_tokens fallback)"
  },
  "truth_label_count": 999,
  "prediction_count": 1000,
  "evaluated_example_count": 999,
  "calibration_metrics": {
    "available": true,
    "sample_count": 961,
    "bin_count": 10,
    "ece": 0.07558792924037576,
    "mce": 0.45,
    "brier_score": 0.1302158168574394
  },
  "first_prompt_timestamp": "2026-03-12T14:43:10.303101Z",
  "last_prompt_timestamp": "2026-03-12T15:37:47.497536Z",
  "overall_time_seconds": 3277.194435,
  "overall_time_human": "54m 37s",
  "accuracy": 0.8228228228228228,
  "macro_precision": 0.666411944957672,
  "macro_recall": 0.6864323966446648,
  "macro_f1": 0.6694183185663711,
  "per_label": {
    "Clause-final pragmatic marker": {
      "precision": 0.6770186335403726,
      "recall": 0.9159663865546218,
      "f1": 0.7785714285714285,
      "support": 119
    },
    "Clause-medial pragmatic marker": {
      "precision": 0.9517241379310345,
      "recall": 0.7263157894736842,
      "f1": 0.8238805970149253,
      "support": 380
    },
    "Discourse marker": {
      "precision": 0.7714987714987716,
      "recall": 0.8509485094850948,
      "f1": 0.8092783505154639,
      "support": 369
    },
    "Quotative marker": {
      "precision": 0.9318181818181818,
      "recall": 0.9389312977099237,
      "f1": 0.9353612167300379,
      "support": 131
    },
    "unclassified": {
      "precision": 0.0,
      "recall": 0.0,
      "f1": 0.0,
      "support": 0
    }
  },
  "confusion_matrix": {
    "Clause-final pragmatic marker": {
      "Clause-final pragmatic marker": 109,
      "Clause-medial pragmatic marker": 1,
      "Discourse marker": 8,
      "Quotative marker": 0,
      "unclassified": 1
    },
    "Clause-medial pragmatic marker": {
      "Clause-final pragmatic marker": 19,
      "Clause-medial pragmatic marker": 276,
      "Discourse marker": 78,
      "Quotative marker": 1,
      "unclassified": 6
    },
    "Discourse marker": {
      "Clause-final pragmatic marker": 32,
      "Clause-medial pragmatic marker": 13,
      "Discourse marker": 314,
      "Quotative marker": 8,
      "unclassified": 2
    },
    "Quotative marker": {
      "Clause-final pragmatic marker": 1,
      "Clause-medial pragmatic marker": 0,
      "Discourse marker": 7,
      "Quotative marker": 123,
      "unclassified": 0
    },
    "unclassified": {
      "Clause-final pragmatic marker": 0,
      "Clause-medial pragmatic marker": 0,
      "Discourse marker": 0,
      "Quotative marker": 0,
      "unclassified": 0
    }
  },
  "labels": [
    "Clause-final pragmatic marker",
    "Clause-medial pragmatic marker",
    "Discourse marker",
    "Quotative marker",
    "unclassified"
  ],
  "total_examples": 999,
  "label_metrics_available": true,
  "run_config": {
    "task_name": "*like* discourse/pragm",
    "prompt_layout": "standard",
    "task_description": "Classification of instance of *like* into four discourse-pragmatic categories: 1. Quotative marker, 2. Discourse marker, 3. Clause-medial pragmatic marker, 4 . Clause-final pragmatic marker",
    "tags": "*like*;discourse;pragmatics;English;v1"
  },
  "source_input_csv": "",
  "source_output_csv": "data/output/like__requesty__anthropicclaudesonnet46__2026-03-12-15-43.csv",
  "source_labels_csv": ""
}