claim
Voting or consensus-based approaches in multi-LLM collaboration mitigate hallucinations and overconfidence by highlighting discrepancies across peer models.

Authors

Sources

Referenced by nodes (1)