LoFa - BoolQ: leaderboard
Metric: Logical Fallacy Resistance after three attack rounds (LFR@3, %): among the questions a model first answers correctly and whose plain false claim it rejects, the share it still answers correctly after up to three rounds of a DeepSeek-R1-written argument using one of ten logical fallacies, mean over the ten fallacy types, three runs; higher is better; 491 BoolQ yes or no items. Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 | 91.08 |
| 2 | DeepSeek V3 | 79.17 |
| 3 | Llama 3.1 405B | 73.61 |
| 4 | GPT-3.5 Turbo | 66.01 |
| 5 | Llama 3.1 70B | 64.7 |
| 6 | Llama 3.1 8B | 62.14 |
Interactive version: theaggregate.ai/benchmark?slug=lofa-boolq · How It Works · Data refreshed daily, snapshot 2026-09-29.