{
  "dataset_name": "LifeStreamingCoT",
  "repo_id": "skyzhou06/LifeStreamingCoT",
  "version": "v0.4.1",
  "created_by": "skyzhou06 with Codex",
  "generation_method": "source_grounded_rule_based_v0.4.1_quality_patch",
  "reasoning_policy": "selective_concise",
  "chunking_method": "semantic_sentence_split_v0.4_refined",
  "refinement_method": "rule_based_quality_patch_v0.4.1",
  "schema": {
    "id": "required",
    "domain": "required",
    "source_dataset": "required",
    "instruction": "required",
    "context": "required",
    "context_chunks": "required",
    "streaming_reasoning": "required",
    "deep_reasoning": "required",
    "answer": "required",
    "response": "required",
    "messages": "required",
    "text": "required",
    "num_chunks": "required",
    "language": "required",
    "split": "required",
    "generation_method": "required",
    "quality_flags": "required",
    "version": "required",
    "reasoning_policy": "required",
    "chunking_method": "required",
    "chunk_labels": "required",
    "skip_chunks": "required",
    "skip_reasons": "required",
    "reasoning_token_budget": "required",
    "original_num_chunks": "required",
    "chunk_split_count": "required",
    "quality_score": "required",
    "is_high_quality": "required",
    "refinement_method": "required",
    "llm_augmented": "required",
    "llm_augmentation_model": "required",
    "rejected_reason": "required",
    "state_tracking_confidence": "required"
  },
  "source_datasets_used": [
    {
      "name": "b-mc2/wikihow_lists",
      "domain": "how_to_guidance",
      "rows": 622
    },
    {
      "name": "pietrolesci/multiwoz_all_versions",
      "domain": "task_oriented_assistant",
      "rows": 2987
    },
    {
      "name": "pixelsandpointers/better_daily_dialog",
      "domain": "daily_dialogue",
      "rows": 3713
    },
    {
      "name": "pixelsandpointers/empathetic_dialogues_for_lm",
      "domain": "emotional_support",
      "rows": 2000
    }
  ],
  "skipped_source_datasets": [],
  "total_rows": 9322,
  "train_rows": 7457,
  "eval_rows": 1865,
  "high_quality_train_rows": 2570,
  "high_quality_eval_rows": 634,
  "domains": {
    "daily_dialogue": 3713,
    "emotional_support": 2000,
    "how_to_guidance": 622,
    "task_oriented_assistant": 2987
  },
  "average_num_chunks": 6.45419437888865,
  "average_chunk_length": 8.952847122959811,
  "average_original_num_chunks": 4.753271830079382,
  "average_chunk_split_count": 1.7806264750053638,
  "average_streaming_reasoning_words": 28.501180004290926,
  "average_deep_reasoning_words": 16.480583565758423,
  "average_quality_score": 0.7602231280841022,
  "high_quality_percentage": 0.3437030680111564,
  "skip_chunk_ratio": 0.11189043645912974,
  "examples_with_at_least_one_skip": 4486,
  "quality_flags_distribution": {
    "awkward_answer": 3441,
    "closing_mishandled": 104,
    "copied_source_response": 469,
    "excessive_chunking": 174,
    "keyword_stitching": 1827,
    "long_deep_reasoning": 2,
    "merged_fragments": 5797,
    "repeated_context_chunks": 537,
    "too_many_skips": 2,
    "weak_context": 10,
    "weak_high_quality_candidate": 3565
  },
  "llm_augmented_count": 0,
  "limitations": [
    "v0.4 is primarily rule-based unless optional LLM augmentation is run.",
    "The high-quality subset is recommended for serious SFT experiments.",
    "Some source datasets are dialogue-style and may not perfectly match ideal assistant behavior.",
    "The dataset is not intended for expert medical, legal, financial, emergency, or safety-critical advice."
  ],
  "samples_for_review": "samples_for_review.jsonl",
  "patch_name": "v0.4.1 loading config and high-quality subset patch",
  "patch_notes": [
    "Adds explicit Hugging Face dataset card configs so default loading uses only data/train.parquet and data/eval.parquet.",
    "Adds a separate high_quality config backed by data/train_high_quality.parquet and data/eval_high_quality.parquet.",
    "Tightens high-quality subset filtering to remove copied-source responses, awkward answer templates, keyword-stitching, repeated chunks, and weak candidates."
  ],
  "hf_config_fixed": true,
  "old_v0_4_counts": {
    "train_rows": 7457,
    "eval_rows": 1865,
    "high_quality_train_rows": 7207,
    "high_quality_eval_rows": 1807,
    "hf_auto_detected_total_rows": 18336
  },
  "high_quality_metrics": {
    "rows": 3204,
    "average_quality_score": 0.9999531835205993,
    "average_streaming_reasoning_words": 34.70536828963795,
    "average_deep_reasoning_words": 21.18196004993758,
    "average_num_chunks": 6.717540574282148,
    "average_chunk_length": 8.440737815360313,
    "skip_chunk_ratio": 0.18882126097662966,
    "severe_flag_percentage": 0.0,
    "quality_flags_distribution": {
      "merged_fragments": 2187,
      "too_many_skips": 2,
      "weak_context": 1
    }
  },
  "high_quality_filtering_rules": [
    "awkward_answer",
    "closing_mishandled",
    "copied_source_response",
    "excessive_chunking",
    "fragment_chunk",
    "generic_reasoning",
    "keyword_stitching",
    "low_specificity",
    "no repeated context chunks",
    "possible_slot_error",
    "quality_score >= 0.85",
    "repeated_context_chunks",
    "streaming/deep length limits",
    "weak_high_quality_candidate"
  ]
}