README.md

---
dataset_info:
- config_name: default
  features:
  - name: id
    dtype: string
  - name: source
    dtype: string
  - name: text
    dtype: string
  - name: category_label
    dtype: int64
  - name: alignment_sentiment
    dtype: string
  - name: is_filter_target
    dtype: bool
  splits:
  - name: train
    num_bytes: 4890940
    num_examples: 536
  download_size: 2926057
  dataset_size: 4890940
- config_name: filter_analysis
  features:
  - name: id
    dtype: string
  - name: source
    dtype: string
  - name: text
    dtype: string
  - name: category
    dtype: string
  - name: category_label
    dtype: int64
  - name: claude_sonnet_predicted_category
    dtype: int64
  - name: claude_haiku_predicted_category
    dtype: int64
  - name: gpt_oss_predicted_category
    dtype: int64
  - name: claude_sonnet_confidence
    dtype: string
  - name: claude_haiku_confidence
    dtype: string
  - name: gpt_oss_confidence
    dtype: string
  - name: claude_sonnet_reasoning
    dtype: string
  - name: claude_haiku_reasoning
    dtype: string
  - name: gpt_oss_reasoning
    dtype: string
  - name: claude_sonnet_intermediate_cot
    dtype: string
  - name: claude_haiku_intermediate_cot
    dtype: string
  - name: claude_sonnet_raw_generation
    dtype: string
  - name: claude_haiku_raw_generation
    dtype: string
  splits:
  - name: llm_judgments_analysis_nov_21_25
    num_bytes: 10194936
    num_examples: 529
  download_size: 5335345
  dataset_size: 10194936
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
- config_name: filter_analysis
  data_files:
  - split: llm_judgments_analysis_nov_21_25
    path: filter_analysis/llm_judgments_analysis_nov_21_25-*
---