ThinknCheck: Grounded Claim Verification with Compact, Reasoning-Driven, and Interpretable Models

TL;DR

ThinknCheck is a compact, reasoning-driven model for grounded claim verification that outperforms larger models and emphasizes interpretability and resource efficiency.

Contribution

It introduces a new training set and fine-tuning approach for a 1B-parameter verifier that produces structured rationales and achieves high accuracy.

Findings

ThinknCheck surpasses larger models on LLMAggreFact with fewer parameters.

Explicit reasoning improves accuracy over zero-shot chain-of-thought.

Domain-specific variants like ThinknCheck-Science enhance performance on specialized benchmarks.

Abstract

We present ThinknCheck, a 1B-parameter verifier for grounded claim verification that first produces a short, structured rationale and then a binary verdict. We construct LLMAggreFact-Think, a 24.1k reasoning-augmented training set derived from LLMAggreFact, and fine-tune a 4-bit Gemma3 model to follow this format. On LLMAggreFact, ThinknCheck attains 78.1 balanced accuracy (BAcc), surpassing MiniCheck-7B (77.4) with 7x fewer parameters; removing the reasoning step reduces BAcc to 57.5. On SciFact, ThinknCheck reaches 64.7 BAcc, a +14.7 absolute gain over MiniCheck-7B. By contrast, zero-shot chain-of-thought on the base Gemma3-1B harms accuracy relative to direct answers, and preference optimization with a simple format+accuracy reward underperforms supervised reasoning. To probe the latter, we introduce GSMClaims and a domain-specialized variant, ThinknCheck-Science, which improves across benchmarks, including 61.0\% accuracy on GSMClaims. Overall, explicit, supervised reasoning enables compact verifiers that are competitive while remaining resource-efficient and interpretable.