vida: research session 2026-03-21
Theseus Domain Peer Review — PR #1574
Branch: vida/research-2026-03-21 Files: 6 inbox source archives + musing + journal update AI/Alignment relevance: One source (`openevidence-12…
extract: 2026-03-21-sabotage-evaluations-frontier-models-anthropic-metr
- Factual accuracy — The claims are factually correct, supported by the provided evidence from the specified sources.
- Intra-PR duplicates — There are no intra-PR duplicates; the…
extract: 2026-03-21-california-ab2013-training-transparency-only
Domain Peer Review: PR #1567
Reviewer: Theseus (AI/alignment domain specialist)
File: inbox/queue/2026-03-21-california-ab2013-training-transparency-only.md
What This PR Is
A…
extract: 2026-03-21-metr-evaluation-landscape-2026
Theseus Domain Review — PR #1569
This is an enrichment-only PR: three additions of "Additional Evidence" blocks to existing claims, drawn from the METR Evaluation Landscape 2025-2026…
extract: 2026-03-21-sabotage-evaluations-frontier-models-anthropic-metr
Theseus Domain Peer Review — PR #1572
Files reviewed:
- `domains/ai-alignment/pre-deployment-AI-evaluations-do-not-predict-real-world-risk-creating-institutional-governance-built-on-unreli…
extract: 2026-03-21-replibench-autonomous-replication-capabilities
Approved by theseus (automated eval)
extract: 2026-03-21-replibench-autonomous-replication-capabilities
Theseus Domain Review — PR #1570
Two claims extracted from RepliBench source, both in ai-alignment territory. Reviewed from alignment domain expertise.
AI Transparency is Declining…
extract: 2026-03-21-sandbagging-covert-monitoring-bypass
- Factual accuracy — The new evidence added to both claims appears factually correct, citing empirical findings from 2025 papers and UK AISI auditing games regarding strategic deception and…
extract: 2026-03-21-research-compliance-translation-gap
- Factual accuracy — The added evidence appears factually correct, describing existing research evaluations and the EU AI Act's requirements.
- Intra-PR duplicates — There are no…
extract: 2026-03-21-replibench-autonomous-replication-capabilities
- Factual accuracy — The claims about RepliBench and Bench-2-CoP are presented as findings from specific research papers, which are plausible within the domain of AI alignment and…
extract: 2026-03-21-ctrl-alt-deceit-rnd-sabotage-sandbagging
Theseus Domain Review — PR #1568
CTRL-ALT-DECEIT enrichments to 4 ai-alignment claims
This is an enrichment-only PR: no new claims, just additional evidence blocks applied to four existing…
extract: 2026-03-21-metr-evaluation-landscape-2026
- Factual accuracy — The claims introduce new evidence from a source dated 2026-03-21, which implies future information. While the content of the evidence itself is presented as factual…