meta_v2.json

{
  "model": "cortyx_v2_final.pt",
  "base": "microsoft/deberta-v3-small",
  "version": "v2.0",
  "epochs": 10,
  "best_f1_macro": 0.6129,
  "labels": [
    "safe",
    "mild_toxicity",
    "severe_toxicity",
    "harassment",
    "hate_speech",
    "threat",
    "insult",
    "profanity",
    "sexual_content",
    "violence",
    "self_harm",
    "extremism",
    "illegal_instruction",
    "jailbreak_attempt",
    "prompt_injection",
    "obfuscated_toxicity",
    "misinformation_risk"
  ],
  "thresholds": {
    "safe": 0.5,
    "mild_toxicity": 0.7,
    "severe_toxicity": 0.4,
    "harassment": 0.5,
    "hate_speech": 0.45,
    "threat": 0.4,
    "insult": 0.55,
    "profanity": 0.6,
    "sexual_content": 0.45,
    "violence": 0.4,
    "self_harm": 0.35,
    "extremism": 0.4,
    "illegal_instruction": 0.45,
    "jailbreak_attempt": 0.45,
    "prompt_injection": 0.45,
    "obfuscated_toxicity": 0.5,
    "misinformation_risk": 0.5
  },
  "datasets": [
    "QuantaSparkLabs/cortyx-safety-dataset",
    "google/civil_comments",
    "lmsys/toxic-chat",
    "lmsys/lmsys-chat-1m",
    "cardiffnlp/tweet_eval"
  ],
  "training_samples": 1844,
  "val_samples": 462
}