Scientific fact-checking is vital for assessing claims in specialized domains such as biomedicine and materials science, yet existing systems often hallucinate or apply inconsistent reasoning, especially when verifying technical, compositional claims against an evidence snippet under source and cost/latency constraints. We present a pipeline centered on atomic predicate-argument decomposition and calibrated, uncertainty-gated corroboration: atomic facts are aligned to local snippets via embeddings, verified by a compact evidence-grounded checker, and only facts with uncertain support trigger domain-restricted web search over authoritative sources. The system supports both binary and tri-valued classification where it predicts labels from Supported, Refuted, NEI for three-way tasks. We evaluate under two regimes, Context-Only (no web) and Context+Web (uncertainty-gated web corroboration); when retrieved evidence conflicts with the provided context, we abstain with NEI rather than overriding the context. On multiple benchmarks, our framework surpasses the strongest benchmarks. In our experiments, web corroboration was invoked for only a minority of atomic facts on average, indicating that external evidence is consulted selectively under calibrated uncertainty rather than routinely. Overall, coupling atomic granularity with calibrated, uncertainty-gated corroboration yields more interpretable and context-conditioned verification, making the approach well-suited to high-stakes, single-document settings that demand traceable rationales, predictable cost/latency, and conservative.
Figures & tables
Atomic+Search
RARR
FacTool
Provenance
Base
Explicit typed atomic facts
✓
✗
✗
✗
✗
Uncertainty-gated retrieval
✓
✗
✗
✗
✗
Domain-restricted external evidence
✓
✗
✗
n/a
n/a
Compact verifier (NLI)
✓
✗
✗
✓
✓
LLM judge aggregation
✓
✗
✓
✗
✗
Post-generation editing objective
✗
✓
✗
✗
✗
Table 1. Comparison of recent fact-checking methods. ✓ indicates the feature is present, ✗ indicates it is absent, and n/a means not applicable. “Base” represents a typical retrieve–verify pipeline.
Figure 1. Atomic+Search pipeline
BIONLI-300
PubMedFact1k (3-way)
CLIMATE-FEVER (2-way)
System
Bal. Acc. ↑
Rec. ↑
F1 ↑
Macro F1 ↑
Bal. Acc. ↑
Rec. (S) ↑
F1 (S) ↑
GPT-o1
65.8%
60.2%
64.9%
71.2%
69.70%
56.10%
65.05%
MiniCheck
61.35%
59.8%
60.7%
—
69.10%
55.00%
64.00%
GPT-5 Mini
62.9%
58.7%
61.8%
68.5%
67.90%
54.20%
63.10%
Qwen 32B MAD
62.5%
59.1%
61.3%
61.8%
67.30%
53.60%
62.60%
RARR
66.4%
62.3%
65.3%
72.3%
70.40%
57.80%
66.30%
Table 2. Main results. Balanced Accuracy, Recall, and F1 on BIONLI-300 ; Macro-F 1 on PubMedFact1k ; and Balanced Accuracy, Recall (Supports), and F1 (Supports) on CLIMATE-FEVER , evaluated only on the Supported / Refuted subsets.
BIONLI-300
F 1 ↑
Δ vs. full
Atomic+Search (full)
66.7%
—
No-Search
62.0%
−4.7
No-Atomic
60.3%
−6.4
MajorityVote (No-Judge)
52.1%
−14.6
Table 3. Ablations on BIONLI-300 . Each line removes one component from the full pipeline.
Figure 2. Distribution of per-fact support pi before and after web corroboration on BIONLI-300 .
Fact-checking systems built on LLMs achieve high verdict accuracy on standard benchmarks, yet routinely output Supports labels whose cited evidence does not license the claim. Structured decomposition is the natural way to inspect those warrants, but rigid extraction protocols strip the full-claim context that facets need. We introduce SIFT -- claim-conditioned re-scoring of extracted evidence spans against the full claim -- paired with WSP (Warranted Supports Proportion), an automatic NLI check that the cited warrant entails the claim. We evaluate on FEVER, SciFact, 5PILS, and DP across four open-source backbones. SIFT recovers accuracy on cells where naive decomposition costs up to 27.6 points, while raising WSP above direct prompting; WSP itself calibrates against human gold evidence at AUC 0.92 and precision 0.98.
Large language models (LLMs) can achieve strong fact-checking accuracy, yet forced binary decisions conceal a critical reliability problem: systems may issue confident verdicts even when supporting evidence is weak, sparse, or internally inconsistent. We address this issue through Evidence Chain Evaluation (ECE), a selective fact-checking framework that permits abstention via an uncertain verdict instead of requiring a true/false decision for every claim. The evaluated system is a tool-using verification agent that gathers evidence through web search, scholarly search, and executable checks, and then returns a structured verdict with confidence and source-level metadata. On ECE-Bench, ECE achieves 91.6% standard accuracy, 93.7% coverage, and 97.8% selective accuracy on answered claims. Although ECE does not outperform the strongest retrieval baseline on aggregate calibration metrics such as Expected Calibration Error, Brier score, or AURC, it delivers a clear selective-prediction trade-off: the system maintains very high accuracy on answered claims while deferring 6 of 95 cases. These deferred cases are concentrated in lower-reliability evidence settings (5/6 at source level L4), supporting the view that abstention functions as a safety-oriented mechanism for handling epistemically weak evidence. Code is available at https://github.com/ cheshireyang/ECE.git
Large language models (LLMs) excel in generating fluent utterances but can lack reliable grounding in verified information. At the same time, knowledge-graph-based fact-checkers deliver precise and interpretable evidence, yet suffer from limited coverage or latency. By integrating LLMs with knowledge graphs and real-time search agents, we introduce a hybrid fact-checking approach that leverages the individual strengths of each component. Our system comprises three autonomous steps: 1) a Knowledge Graph (KG) Retrieval for rapid one-hop lookups in DBpedia, 2) an LM-based classification guided by a task-specific labeling prompt, producing outputs with internal rule-based logic, and 3) a Web Search Agent invoked only when KG coverage is insufficient. Our pipeline achieves an F1 score of 0.93 on the FEVER benchmark on the Supported/Refuted split without task-specific fine-tuning. To address Not enough information cases, we conduct a targeted reannotation study showing that our approach frequently uncovers valid evidence for claims originally labeled as Not Enough Information (NEI), as confirmed by both expert annotators and LLM reviewers. With this paper, we present a modular, opensource fact-checking pipeline with fallback strategies and generalization across datasets.
Shaghayegh Kolli, Richard Rosenbaum, Timo Cavelius +3