AI-Generated Text Detection on KatFish corpus 2025 (Phase 1)
96.67AccuracyGPT-5.2 Thinking
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-5.2 ThinkingEvaluator Category=LLMs, Mode=Zero-shot, Protocol=Blind classification2026.01 | 96.67 | |
| Gemini 3 FlashEvaluator Category=LLMs, Mode=Zero-shot, Protocol=Blind classification2026.01 | 96.67 | |
| LLMs Majority VoteEvaluator Category=LLMs, Aggregation=Majority Vote, Mode=Zero-shot2026.01 | 96.67 | |
| Claude Sonnet 4.5Evaluator Category=LLMs, Mode=Zero-shot, Protocol=Blind classification2026.01 | 90 | |
| Annotator 3Evaluator Category=Human, Decision Criteria=Intuition-driven, Protocol=Blind classification2026.01 | 80 | |
| Human Majority VoteEvaluator Category=Human, Aggregation=Majority Vote, Decision Criteria=Intuition-driven2026.01 | 60 | |
| Annotator 1Evaluator Category=Human, Decision Criteria=Intuition-driven, Protocol=Blind classification2026.01 | 50 | |
| Annotator 2Evaluator Category=Human, Decision Criteria=Intuition-driven, Protocol=Blind classification2026.01 | 46.67 |