Tonic's picture
Update README.md
48cabb9 verified
metadata
dataset_info:
  features:
    - name: label
      dtype: string
    - name: identifier
      dtype: string
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 71882488
      num_examples: 82603
    - name: test
      num_bytes: 405966
      num_examples: 1219
  download_size: 31200331
  dataset_size: 72288454
source_datasets:
  - Tonic/climate-guard-thinking_data_nocomment_qwen_toxic_agent
  - Tonic/climate-guard-synthetic_data_qwen_toxic_agent
  - Tonic/climate-guard-thinking_data_nocomment_intern_toxic_agent
  - Tonic/climate-guard-thinking_data_nocomment_phi4_toxic_agent
  - Tonic/climate-guard-thinking_data_nocomment_yi_toxic_agent
  - Tonic/climate-guard-synthetic_data_nocomment_yi_toxic_agent
  - climate_fever
  - QuotaClimat/frugalaichallenge-text-train
  - takara-ai/QuotaClimat
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
license: mit
task_categories:
  - text-classification
  - zero-shot-classification
language:
  - en
tags:
  - climate
pretty_name: Climate Guard Toxic Agent
size_categories:
  - 10K<n<100K

Climate Guard - Toxic Agent - Dataset

Dataset Description

This dataset is a comprehensive consolidation of multiple climate-related datasets, focusing on climate disinformation and factual climate information. It combines and standardizes data from various high-quality sources to create a robust resource for climate-related text classification tasks.

Dataset Sources

The dataset incorporates data from the following sources:

Data Processing - 👇🏻Click to expand

The dataset underwent several processing steps to ensure quality and consistency:

  1. Text Cleaning:

    • Removed responses starting with apology phrases ("I'm sorry", "I am sorry", "I apologize", "Given the directive")
    • Cleaned text between "---" markers
    • Standardized text formatting
  2. Label Standardization:

    • Maintained consistent label format across all sources
    • Special handling for '0_not_relevant' labels from specific sources
  3. Source Tracking:

    • Added source identifiers to track data origin
    • Preserved dataset provenance information

Dataset Structure

The dataset is split into training and testing sets with the following features:

DatasetDict({
    'train': Dataset({
        features: ['identifier', 'text', 'label'],
        num_examples: <num_examples>
    }),
    'test': Dataset({
        features: ['identifier', 'text', 'label'],
        num_examples: <num_examples>
    })
})

Features:

  • identifier: String identifying the source dataset
  • text: The main text content
  • label: Classification label

Labels:

  • 0_not_relevant: No relevant claim detected or claims that don't fit other categories
  • 1_not_happening: Claims denying the occurrence of global warming and its effects - Global warming is not happening. Climate change is NOT leading to melting ice (such as glaciers, sea ice, and permafrost), increased extreme weather, or rising sea levels. Cold weather also shows that climate change is not happening
  • 2_not_human: Claims denying human responsibility in climate change - Greenhouse gases from humans are not the causing climate change.
  • 3_not_bad: Claims minimizing or denying negative impacts of climate change - The impacts of climate change will not be bad and might even be beneficial.
  • 4_solutions_harmful_unnecessary: Claims against climate solutions - Climate solutions are harmful or unnecessary
  • 5_science_is_unreliable: Claims questioning climate science validity - Climate science is uncertain, unsound, unreliable, or biased.
  • 6_proponents_biased: Claims attacking climate scientists and activists - Climate scientists and proponents of climate action are alarmist, biased, wrong, hypocritical, corrupt, and/or politically motivated.
  • 7_fossil_fuels_needed: Claims promoting fossil fuel necessity - We need fossil fuels for economic growth, prosperity, and to maintain our standard of living.

Label Distribution Train

Label Source Heatmap Train

Source Distribution Train

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("Tonic/Climate-Guard-Toxic-Agent")

# Access training data
train_data = dataset['train']

# Access test data
test_data = dataset['test']

Dataset Statistics - 👇🏻Click to expand

{
  "basic_stats": {
    "total_samples": {
      "train": 82603,
      "test": 1219
    },
    "label_distribution": {
      "train": {
        "3_not_bad": 11011,
        "4_solutions_harmful_unnecessary": 11597,
        "5_science_is_unreliable": 14609,
        "6_proponents_biased": 8494,
        "7_fossil_fuels_needed": 10585,
        "1_not_happening": 11380,
        "2_not_human": 11772,
        "0_not_relevant": 3155
      },
      "test": {
        "6_proponents_biased": 139,
        "2_not_human": 137,
        "3_not_bad": 97,
        "1_not_happening": 154,
        "5_science_unreliable": 160,
        "4_solutions_harmful_unnecessary": 160,
        "7_fossil_fuels_needed": 65,
        "0_not_relevant": 307
      }
    },
    "source_distribution": {
      "train": {
        "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1546,
        "climate-guard-synthetic_data_qwen_toxic_agent": 32209,
        "climate-guard-thinking_data_nocomment_intern_toxic_agent": 3297,
        "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 28510,
        "climate-guard-thinking_data_nocomment_yi_toxic_agent": 1687,
        "climate-guard-synthetic_data_yi_toxic_agent": 3789,
        "frugal_challenge_train": 1311,
        "climate_fever": 654,
        "quota_climat": 9600
      },
      "test": {
        "frugal_challenge_test": 1219
      }
    },
    "label_source_incidence": {
      "train": {
        "counts": {
          "3_not_bad": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 207,
            "climate-guard-synthetic_data_qwen_toxic_agent": 4116,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 478,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 4332,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 218,
            "climate-guard-synthetic_data_yi_toxic_agent": 496,
            "quota_climat": 1164
          },
          "4_solutions_harmful_unnecessary": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 223,
            "climate-guard-synthetic_data_qwen_toxic_agent": 4760,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 473,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 4112,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 236,
            "climate-guard-synthetic_data_yi_toxic_agent": 557,
            "quota_climat": 1236
          },
          "5_science_is_unreliable": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 300,
            "climate-guard-synthetic_data_qwen_toxic_agent": 5454,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 604,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 6091,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 318,
            "climate-guard-synthetic_data_yi_toxic_agent": 656,
            "quota_climat": 1186
          },
          "6_proponents_biased": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 167,
            "climate-guard-synthetic_data_qwen_toxic_agent": 4535,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 389,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 1389,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 234,
            "climate-guard-synthetic_data_yi_toxic_agent": 544,
            "quota_climat": 1236
          },
          "7_fossil_fuels_needed": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 205,
            "climate-guard-synthetic_data_qwen_toxic_agent": 3979,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 424,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 4143,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 217,
            "climate-guard-synthetic_data_yi_toxic_agent": 476,
            "quota_climat": 1141
          },
          "1_not_happening": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 227,
            "climate-guard-synthetic_data_qwen_toxic_agent": 4700,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 466,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 3976,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 236,
            "climate-guard-synthetic_data_yi_toxic_agent": 548,
            "quota_climat": 1227
          },
          "2_not_human": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 217,
            "climate-guard-synthetic_data_qwen_toxic_agent": 4665,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 463,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 4467,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 228,
            "climate-guard-synthetic_data_yi_toxic_agent": 512,
            "quota_climat": 1220
          },
          "0_not_relevant": {
            "frugal_challenge_train": 1311,
            "climate_fever": 654,
            "quota_climat": 1190
          }
        },
        "percentages": {
          "3_not_bad": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.879938243574607,
            "climate-guard-synthetic_data_qwen_toxic_agent": 37.38080101716466,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.3411134320225235,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 39.34247570611207,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 1.979838343474707,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.504586322768141,
            "quota_climat": 10.571246934883298
          },
          "4_solutions_harmful_unnecessary": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.9229110976976806,
            "climate-guard-synthetic_data_qwen_toxic_agent": 41.04509787013883,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.07864102785203,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 35.45744589117875,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 2.0350090540657066,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.802966284383892,
            "quota_climat": 10.657928774683107
          },
          "5_science_is_unreliable": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 2.053528646724622,
            "climate-guard-synthetic_data_qwen_toxic_agent": 37.33315079745363,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.134437675405572,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 41.693476623998905,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 2.176740365528099,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.490382640837839,
            "quota_climat": 8.118283250051338
          },
          "6_proponents_biased": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.9660937132093244,
            "climate-guard-synthetic_data_qwen_toxic_agent": 53.390628679067575,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.579703319990582,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 16.352719566753002,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 2.7548858017424065,
            "climate-guard-synthetic_data_yi_toxic_agent": 6.404520838238757,
            "quota_climat": 14.551448080998352
          },
          "7_fossil_fuels_needed": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.9367028814359941,
            "climate-guard-synthetic_data_qwen_toxic_agent": 37.590930562116206,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.005668398677374,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 39.140292867264996,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 2.050070854983467,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.496929617383089,
            "quota_climat": 10.779404818138875
          },
          "1_not_happening": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.9947275922671353,
            "climate-guard-synthetic_data_qwen_toxic_agent": 41.30052724077329,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 4.094903339191564,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 34.93848857644991,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 2.0738137082601056,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.815465729349736,
            "quota_climat": 10.782073813708259
          },
          "2_not_human": {
            "climate-guard-thinking_data_nocomment_qwen_toxic_agent": 1.8433571185864763,
            "climate-guard-synthetic_data_qwen_toxic_agent": 39.627930682976555,
            "climate-guard-thinking_data_nocomment_intern_toxic_agent": 3.933061501868841,
            "climate-guard-thinking_data_nocomment_phi4_toxic_agent": 37.94597349643221,
            "climate-guard-thinking_data_nocomment_yi_toxic_agent": 1.9367991845056065,
            "climate-guard-synthetic_data_yi_toxic_agent": 4.349303431872239,
            "quota_climat": 10.36357458375807
          },
          "0_not_relevant": {
            "frugal_challenge_train": 41.55309033280507,
            "climate_fever": 20.72900158478605,
            "quota_climat": 37.717908082408876
          }
        }
      },
      "test": {
        "counts": {
          "6_proponents_biased": {
            "frugal_challenge_test": 139
          },
          "2_not_human": {
            "frugal_challenge_test": 137
          },
          "3_not_bad": {
            "frugal_challenge_test": 97
          },
          "1_not_happening": {
            "frugal_challenge_test": 154
          },
          "5_science_unreliable": {
            "frugal_challenge_test": 160
          },
          "4_solutions_harmful_unnecessary": {
            "frugal_challenge_test": 160
          },
          "7_fossil_fuels_needed": {
            "frugal_challenge_test": 65
          },
          "0_not_relevant": {
            "frugal_challenge_test": 307
          }
        },
        "percentages": {
          "6_proponents_biased": {
            "frugal_challenge_test": 100.0
          },
          "2_not_human": {
            "frugal_challenge_test": 100.0
          },
          "3_not_bad": {
            "frugal_challenge_test": 100.0
          },
          "1_not_happening": {
            "frugal_challenge_test": 100.0
          },
          "5_science_unreliable": {
            "frugal_challenge_test": 100.0
          },
          "4_solutions_harmful_unnecessary": {
            "frugal_challenge_test": 100.0
          },
          "7_fossil_fuels_needed": {
            "frugal_challenge_test": 100.0
          },
          "0_not_relevant": {
            "frugal_challenge_test": 100.0
          }
        }
      }
    }
  },
  "text_stats": {
    "train": {
      "avg_length": 111.46446254978633,
      "median_length": 77.0,
      "std_length": 114.89517560291323,
      "min_length": 0,
      "max_length": 965,
      "total_words": 9207299
    },
    "test": {
      "avg_length": 46.73502871205906,
      "median_length": 37.0,
      "std_length": 37.74882897285664,
      "min_length": 4,
      "max_length": 454,
      "total_words": 56970
    }
  },
  "vocabulary_stats": {
    "train": {
      "vocabulary_size": 70216,
      "total_tokens": 9207299,
      "unique_tokens_ratio": 0.007626123578695554
    },
    "test": {
      "vocabulary_size": 10676,
      "total_tokens": 56970,
      "unique_tokens_ratio": 0.18739687554853432
    }
  },
  "label_patterns": {
    "train": {
      "dominant_sources_per_label": {
        "3_not_bad": {
          "main_source": "climate-guard-thinking_data_nocomment_phi4_toxic_agent",
          "percentage": 39.34247570611207
        },
        "4_solutions_harmful_unnecessary": {
          "main_source": "climate-guard-synthetic_data_qwen_toxic_agent",
          "percentage": 41.04509787013883
        },
        "5_science_is_unreliable": {
          "main_source": "climate-guard-thinking_data_nocomment_phi4_toxic_agent",
          "percentage": 41.693476623998905
        },
        "6_proponents_biased": {
          "main_source": "climate-guard-synthetic_data_qwen_toxic_agent",
          "percentage": 53.390628679067575
        },
        "7_fossil_fuels_needed": {
          "main_source": "climate-guard-thinking_data_nocomment_phi4_toxic_agent",
          "percentage": 39.140292867264996
        },
        "1_not_happening": {
          "main_source": "climate-guard-synthetic_data_qwen_toxic_agent",
          "percentage": 41.30052724077329
        },
        "2_not_human": {
          "main_source": "climate-guard-synthetic_data_qwen_toxic_agent",
          "percentage": 39.627930682976555
        },
        "0_not_relevant": {
          "main_source": "frugal_challenge_train",
          "percentage": 41.55309033280507
        }
      },
      "label_diversity_per_source": {
        "climate-guard-thinking_data_nocomment_qwen_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.9322811905174009
        },
        "climate-guard-synthetic_data_qwen_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.9412569930894747
        },
        "climate-guard-thinking_data_nocomment_intern_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.9376010166020219
        },
        "climate-guard-thinking_data_nocomment_phi4_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.8879859048798708
        },
        "climate-guard-thinking_data_nocomment_yi_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.9375508611483394
        },
        "climate-guard-synthetic_data_yi_toxic_agent": {
          "unique_labels": 7,
          "entropy": 1.941023858626436
        },
        "frugal_challenge_train": {
          "unique_labels": 1,
          "entropy": 0.0
        },
        "climate_fever": {
          "unique_labels": 1,
          "entropy": 0.0
        },
        "quota_climat": {
          "unique_labels": 8,
          "entropy": 2.0790581410796753
        }
      },
      "source_bias_analysis": {
        "climate-guard-thinking_data_nocomment_qwen_toxic_agent": {
          "kl_divergence": 0.0395716559443841
        },
        "climate-guard-synthetic_data_qwen_toxic_agent": {
          "kl_divergence": 0.045281501914864145
        },
        "climate-guard-thinking_data_nocomment_intern_toxic_agent": {
          "kl_divergence": 0.039965634146765544
        },
        "climate-guard-thinking_data_nocomment_phi4_toxic_agent": {
          "kl_divergence": 0.06259067672088119
        },
        "climate-guard-thinking_data_nocomment_yi_toxic_agent": {
          "kl_divergence": 0.044481091436281824
        },
        "climate-guard-synthetic_data_yi_toxic_agent": {
          "kl_divergence": 0.04597417615615136
        },
        "frugal_challenge_train": {
          "kl_divergence": 3.265057483962074
        },
        "climate_fever": {
          "kl_divergence": 3.265057483962074
        },
        "quota_climat": {
          "kl_divergence": 0.07482175184545027
        }
      }
    },
    "test": {
      "dominant_sources_per_label": {
        "6_proponents_biased": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "2_not_human": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "3_not_bad": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "1_not_happening": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "5_science_unreliable": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "4_solutions_harmful_unnecessary": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "7_fossil_fuels_needed": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        },
        "0_not_relevant": {
          "main_source": "frugal_challenge_test",
          "percentage": 100.0
        }
      },
      "label_diversity_per_source": {
        "frugal_challenge_test": {
          "unique_labels": 8,
          "entropy": 1.9926606322233085
        }
      },
      "source_bias_analysis": {
        "frugal_challenge_test": {
          "kl_divergence": 0.0
        }
      }
    }
  }
}

Intended Uses

This dataset is designed for:

  • Training climate disinformation detection models
  • Developing fact-checking systems
  • Analyzing climate-related discourse patterns
  • Research in climate communication

Limitations

  • The dataset may contain some inherent biases from source datasets
  • Some sources are synthetic or AI-generated data
  • Language is primarily English
  • Coverage may vary across different types of climate disinformation

Citation

If you use this dataset, please cite both this consolidated version and the original source datasets:

@dataset{consolidated_climate_dataset,
    author = {Joseph Pollack},
    title = {Climate Guard Toxic Agent - Dataset},
    year = {2024},
    publisher = {Hugging Face},
    url = {https://huggingface.co/datasets/Tonic/Climate-Guard-Toxic-Agent}
}
@article{coan2021computer,
  title={Computer-assisted classification of contrarian claims about climate change},
  author={Coan, Travis G and Boussalis, Constantine and Cook, John and others},
  journal={Scientific Reports},
  volume={11},
  number={22320},
  year={2021},
  publisher={Nature Publishing Group},
  doi={10.1038/s41598-021-01714-4}
}

License

This dataset is released under the same licenses as its source datasets. Please refer to individual source datasets for specific license information.

Contact

For questions or issues regarding this dataset, please open a community comment or descriptive PR.

Acknowledgments

We thank the creators and maintainers of all source datasets used in this consolidation:

  • The Tonic AI team
  • QuotaClimat team
  • Climate FEVER dataset creators
  • Takara AI team
  • All other contributors to the source datasets

Updates and Maintenance

This dataset will be periodically updated to:

  • Fix any identified issues
  • Include new relevant source datasets
  • Improve data quality and consistency

Last updated: [Current Date] Version: 1.0.0