Using LLaMA-3.1-70B as the sole evaluation model in the HalluLens benchmark raises concerns about bias, particularly when the benchmark is used to judge other LLaMA variants.

Authors

• Person: Pascale Fung Organization: LinkedIn
  Pascale Fung's Post - LLM Hallucination Benchmark

Sources

• Pascale Fung's Post - LLM Hallucination Benchmark www.linkedin.com Pascale Fung · LinkedIn via serper

Referenced by nodes (1)

• LLaMA concept