Safeguard-Conditioned Uplift：双用途生物学助手的效用-风险前沿评估框架

View PDF HTML (experimental)

Abstract:Refusal rates conflate upstream provider behavior with downstream safeguards and do not reveal what information remains in a delivered answer. Safeguard-conditioned uplift is an evidence-gated framework for evaluating benchmark utility and risk-related proxies along deployed access paths for dual-use biology assistants. It attributes interventions to their source, selects thresholds under a legitimate-access budget, and supports content claims only when answer verifiers pass frozen qualification tests. In a prospectively frozen route and generation study, two of six configurations pass the LAB-Bench-calibrated action-level rule, but both inherit the same Claude Opus provider effect rather than independent downstream gains. Positive selectivity persists on 104 previously unused matched pairs, but benign intervention exceeds the $20\%$ limit. No joint-scoring verifier meets the frozen criteria; a factorial follow-up shows that criterion isolation and ordinal outputs improve end-to-end edit-direction accuracy on controlled template edits. The evidence supports action-level selectivity and verifier interface effects, but not calibrated access, verified content removal, or biological-risk reduction.

Comments:	30 pages, 3 figures, 36 tables
Subjects:	Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
Cite as:	arXiv:2607.13039 [cs.CY]
	(or arXiv:2607.13039v3 [cs.CY] for this version)
	https://doi.org/10.48550/arXiv.2607.13039 arXiv-issued DOI via DataCite

Submission history

From: Dipesh Tharu Mahato [view email]
[v1] Fri, 12 Jun 2026 00:57:14 UTC (65 KB)
[v2] Mon, 17 Aug 2026 03:05:07 UTC (108 KB)
[v3] Sun, 25 Oct 2026 23:17:17 UTC (124 KB)