LoFa - NQ2: 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; 489 Natural Questions multiple-choice items (second subset). Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.

Top models

#ModelScore
1DeepSeek V393.97
2GPT-490.75
3Llama 3.1 405B83.33
4Llama 3.1 70B80.08
5GPT-3.5 Turbo73.44
6Llama 3.1 8B73.38

Interactive version: theaggregate.ai/benchmark?slug=lofa-nq2 · How It Works · Data refreshed daily, snapshot 2026-09-29.