LexAgentHallu - LRAS - Overall Hallucination Frequency: leaderboard
LexAgentHallu evaluates hallucinations in multi-step legal agents. L1 covers substantive legal knowledge and L2 procedural reasoning. Hallucination frequency is the fraction of rollouts with at least one flagged error; density is the mean fraction of applicable error subclasses triggered. Frameworks are separate protocols; the LRAS fine-tuned models retain their own identities. This series reports LexAgentHallu - LRAS - Overall Hallucination Frequency from 2609.09754v1, S4.T2: Hallucination Frequency (%). Lower is better.
Metric: Hallucination Frequency (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | LRAS-Qwen3-14B | 91.1 | |
| 2 | LRAS-Qwen3-8B | 91.4 | |
| 3 | LRAS-Qwen3-4B | 92.4 |
Interactive version: theaggregate.ai/benchmark?slug=lexagenthallu-lras-overall-hallucination-frequency · How It Works · Data refreshed daily, snapshot 2026-11-08.