PPT-Bench (Epistemic Attack) - Authority Inversion: leaderboard
Metric: Binary capitulation rate (%): share of items scored 1 or 2 over the 22 Type 3 items (authority inversion: epistemic anarchism, anti-foundationalism, radical empiricism); a gpt-4o pairwise judge compares each single-turn pressured answer (L1) with the model's own unpressured baseline answer (L0) and scores 0 (held), 1 (partial) or 2 (full capitulation); temperature 0, one generation per item; lower is better. Source: arxiv.org. Saturation forecast: Around January 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-OSS-20B | 27.3 |
| 2 | GPT-OSS-20B (Low) | 31.8 |
| 3 | Nemotron 3 Super | 31.8 |
| 4 | Mistral Small 3.1 | 50 |
| 5 | Qwen 3 32B | 68.2 |
| 6 | Llama 3.3 70B Instruct | 72.7 |
| 7 | DeepSeek V3.1 | 77.3 |
| 8 | Ministral 8B | 86.4 |
Interactive version: theaggregate.ai/benchmark?slug=ppt-bench-epistemic-attack-authority-inversion · How It Works · Data refreshed daily, snapshot 2026-10-07.