Evidence item

The abstract reports CausalVul’s Devign F1 score and improvements over CodeBERT and UniXcoder under group-aware evaluation. MEASUREMENT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 14759

Source: Causal-aware training with counterfactual supervision and structural causal models improves the robustness of LLM-based vulnerability detection · locator {} · excerpt hash sha256:4d9b9eed2023e6f0e3d6de3ecce443f7c292c4e69e89ec3f126b8c3cf43c9787

Exact excerpt

Under a rigorous project-level group-aware evaluation protocol, CausalVul achieves an F1-score of 76.4% on Devign—an improvement of +18.0 points over its own fine-tuned CodeBERT backbone under identical conditions, and +4.8 points over the strongest cross-backbone baseline (UniXcoder, 71.6%)—with an AUC-ROC of 0.862 and a Mechanism-Grounded Accuracy (the joint accuracy of the detection and CWE heads) of 54.8%.

Linked claims

Claim	Direction	Relevance	Steps	Counted
330f1be0	SUPPORT	DIRECT	0	yes

Audits

None.