Llama-3.1-8B

'Llama 3.1 8B' is a benchmark dataset used to evaluate the effectiveness of methods for detecting tool use hallucinations in autonomous agents through spectral analysis of attention topology.

Papers using Llama-3.1-8B (2)

Blind Spots in the Guard: How Domain-Camouflaged Injection Attacks Evade Detection in Multi-Agent LLM Systems2026Spectral Guardrails for Agents in the Wild: Detecting Tool Use Hallucinations via Attention Topology2026