Fine-tuning LLMs on 350 atomic reasoning examples improves scientific table claim verification, but the claimed edge over GPT-4o CoT only appears on the public SciTab benchmark.
The data in the table must clearly support the claim based on physical principles of the domain or experimental facts, not just superficial correlations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Atomic Reasoning for Scientific Table Claim Verification
Fine-tuning LLMs on 350 atomic reasoning examples improves scientific table claim verification, but the claimed edge over GPT-4o CoT only appears on the public SciTab benchmark.