{"slug": "improving-llms-via-validator-to-generator-alignment", "title": "Improving LLMs via Validator-to-Generator Alignment", "summary": "Researchers introduced a new method, FCPA, to align large language models' generators with their validators, improving consistency by correcting for utterance frequency. Training with FCPA boosted Pearson correlation by up to 27 percentage points on IFEval and HumanEval while preserving validator quality.", "body_md": "arXiv:2607.02668v1 Announce Type: new\nAbstract: Large language models are inconsistent: varying prompts or including unrelated information can lead to unexpected changes in model outputs. The generator-validator (G-V) gap is one manifestation of this phenomenon, where LLMs generate responses that they then deem as invalid if re-queried to validate them. In this work, we introduce a new formulation of G-V consistency that involves a principled correction for utterance frequency. Specifically, generators often assign low likelihood to valid strings simply because those strings are a priori unlikely, which makes naive notions of G-V consistency unworkable. We show that under a natural model of rational agents answering questions with multiple answers, consistency of the validator with a frequency-corrected generator score emerges naturally. Our method, \\emph{\\FCPAname} (\\FCPA), is a training objective implementing frequency-corrected G-V consistency for real-world LLMs. Our experimental results show that training with \\FCPA{} substantially improves both G-V consistency and generator performance over prior methods, with gains of up to $+27$pp in Pearson correlation on IFEval and HumanEval, while preserving validator quality across all evaluated tasks.", "url": "https://wpnews.pro/news/improving-llms-via-validator-to-generator-alignment", "canonical_source": "https://arxiv.org/abs/2607.02668", "published_at": "2026-07-07 04:00:00+00:00", "updated_at": "2026-07-07 04:01:18.890069+00:00", "lang": "en", "topics": ["large-language-models", "ai-research"], "entities": ["arXiv"], "alternates": {"html": "https://wpnews.pro/news/improving-llms-via-validator-to-generator-alignment", "markdown": "https://wpnews.pro/news/improving-llms-via-validator-to-generator-alignment.md", "text": "https://wpnews.pro/news/improving-llms-via-validator-to-generator-alignment.txt", "jsonld": "https://wpnews.pro/news/improving-llms-via-validator-to-generator-alignment.jsonld"}}