← all papers · overview

Truthflow: Truthful LLM Generation Via Representation Flow Correction

Abstract

Large language models (LLMs) are known to struggle with consistently generating truthful responses. While various representation intervention techniques have been proposed, these methods typically apply a universal representation correction vector to all input queries, limiting their effectiveness against diverse queries in practice. In this study, we introduce TruthFlow, a novel method that lever

Related papers

Ranked by semantic similarity — how closely each paper's abstract matches this one (100% = near-identical topic).