KLCF formalizes long-form factuality as bidirectional distribution matching between expressed and parametric knowledge, using a sampled factual checklist for recall and a truthfulness reward for precision.
Objects” subset. For the Factory benchmark, we randomly select 250 samples from its “Hard
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Knowledge-Level Consistency Reinforcement Learning: Dual-Fact Alignment for Long-Form Factuality
KLCF formalizes long-form factuality as bidirectional distribution matching between expressed and parametric knowledge, using a sampled factual checklist for recall and a truthfulness reward for precision.