Evaluator Bench

Sources

122 sources cited by signals: 109 confirmed, 1 imported, 10 unaudited, 2 unverifiable. Tier 1 is a filing or a funder's own index; tier 2 a third-party ledger; tier 3 the organization's own statement; tier 4 press.

SourcePublisherPublishedTierAuditSignals
It looks like there are some good funding opportunities in AI safety right now80,000 Hours2025-10-13tier 4 pressconfirmed3
AI Evaluator Forum launch and AEF-1AI Evaluator Forum2025-12-04tier 3 selfconfirmed6
Evaluation Transparency LetterAI Evaluator Forum2026tier 3 selfconfirmed1
UK AI Security InstituteAI Wiki2026-05-07tier 4 pressconfirmed1
Illinois SB 315: A State Strategy for Enduring National AI Safety StandardsAkerman LLP2026-06-10tier 4 pressconfirmed1
METR raises $71M to independently stress-test AIAlphaSignal2026-08tier 4 pressconfirmed1
Andon LabsAndon Labs2026-07tier 3 selfconfirmed7
AboutApollo Research2026tier 3 selfconfirmed4
Apollo Research is becoming a PBCApollo Research2026-01-20tier 3 selfconfirmed1
Apollo Research is becoming a PBC (governance)Apollo Research2026-01-20tier 3 selfconfirmed1
Apollo Research is becoming a PBC (investment section)Apollo Research2026-01-20tier 3 selfconfirmed3
Our Norms on Security, Science Communication and Conflicts of InterestApollo Research2025-11-26tier 3 selfconfirmed4
The First Year of Apollo ResearchApollo Research2024tier 3 selfconfirmed1
Expanding External Access to Frontier AI Models for Dangerous Capability Evaluations (arXiv 2601.11916)arXiv2026-01-21tier 4 pressconfirmed1
Frontier AI Auditing: Toward Rigorous Third-Party Assessment (arXiv 2601.11699)arXiv2026-02-07tier 4 pressconfirmed1
AboutAVERI2026tier 3 selfconfirmed4
About (funding and conflicts)AVERI2026tier 3 selfconfirmed4
AVERI Pilot Report: The World's First Double-Blind Evaluation of a Proprietary Language ModelAVERI2026-08tier 3 selfconfirmed6
Center for AI SafetyCAIS2026tier 3 selfconfirmed7
CAISI Frontier Testing Agreements Reach Five LabsCloud Security Alliance2026-05-20tier 4 pressunaudited5
UK AISI's Frontier AI Trends Report: Security ImplicationsCloud Security Alliance2026-07-12tier 4 pressconfirmed7
CAISI Frontier Testing Agreements Reach Five LabsCloud Security Alliance (research note)2026-05-05tier 4 pressconfirmed2
Scale AI not winding down following Meta deal, interim CEO saysCNBC2025-06-18tier 4 pressconfirmed1
Trump's head of AI safety agency CAISI resigns after months on jobCNBC2026-07-20tier 4 pressconfirmed1
Coefficient Giving grants indexCoefficient Giving2026-09-11tier 1 indeximported8
Coefficient Giving homeCoefficient Giving2026tier 3 selfconfirmed1
Illinois AI Safety Measures Act SB 315Crowell & Moring2026-07-08tier 4 pressconfirmed3
Illinois Enacts Frontier AI Safety LawDavis Wright Tremaine2026-07-15tier 3 selfconfirmed1
Gray Swan raises $40M Series A to secure frontier AIDealroom2026-06tier 4 pressunaudited1
Clarifying the creation and use of the FrontierMath benchmarkEpoch AI2025-01-23tier 3 selfconfirmed5
TransparencyEpoch AI2026tier 3 selfconfirmed3
TransparencyEpoch AI2026tier 3 selfconfirmed2
FAR.AI Secures Over $30 Million in Multi-Funder SupportFAR.AI2026-07-16tier 3 selfconfirmed6
TransparencyFAR.AI2026tier 3 selfconfirmed1
Gray Swan AI Is Working With OpenAI to Red Team Its ModelsForbes2024-10-29tier 4 pressconfirmed1
This AI Startup's Army of 15,000 HackersForbes2026-05-28tier 4 pressconfirmed3
Former OpenAI policy chief debuts new institute called AVERIFortune2026-01-15tier 4 pressconfirmed1
Hugging Face goes from a scrappy startup to $13 billion Nvidia acquisitionFortune2026-09-03tier 4 pressconfirmed1
Frontier AI grantmakingFounders Pledge2026tier 3 selfconfirmed1
Gemini model documentation, external testingGoogle DeepMind2025tier 3 selfconfirmed8
Gray Swan announces Series AGray Swan2026-05-28tier 3 selfconfirmed6
Humane IntelligenceHumane Intelligence2026tier 3 selfconfirmed8
Who Funds the AI Safety WatchdogsInside The Black Box (Substack)2026-06-03tier 4 pressconfirmed8
Funding for CAISIInstitute for Progress2026tier 4 pressconfirmed2
What Will It Cost for the US to Be Ready for the Next Big AI Breakthrough?Institute for Progress2026tier 4 pressconfirmed1
Alignment Research Center 990 reportInstrumentl2025tier 2 ledgerunaudited1
METR donor rule wording over time (Wayback captures of metr.org/about and /donate)Internet Archive2025-08tier 2 ledgerunverifiable1
evaluators.csv (24 evaluators: Coefficient and SFF totals, lab money, government contracts, leadership)Kevin Bass (GitHub)2026-09-14tier 2 ledgerunverifiable28
metr-money-figure research ledger and auditsKevin Bass (GitHub)2026-09-14tier 2 ledgerconfirmed4
Illinois Joins Growing State-Level EffortLatham & Watkins2026-07-15tier 4 pressconfirmed1
FAR AILongterm Wiki2026tier 2 ledgerunaudited1
FAR AILongterm Wiki2026-02-26tier 4 pressconfirmed7
Third-Party Model AuditingLongterm Wiki2026-01-29tier 4 pressconfirmed4
Why We Are Co-Leading Gray Swan's Series AMadrona2026-05-28tier 4 pressconfirmed1
Apollo Research on ManifundManifund2024tier 3 selfconfirmed1
Transluce: Fund Scalable Democratic Oversight of AIManifund2026tier 3 selfconfirmed5
Anthropic's 3-Step 'Pace the Frontier' PlanMarkTechPost2026-09-13tier 3 selfconfirmed2
About METRMETR2026-08tier 3 selfconfirmed6
Brief independent investigation of agents' behavior in the OpenAI / Hugging Face hacking incidentMETR2026-08-26tier 3 selfconfirmed7
Frontier AI safety regulations: A reference for lab staffMETR2026-01-29tier 3 selfconfirmed7
Funding updateMETR2026-08-14tier 3 selfconfirmed2
New Support Through The Audacious ProjectMETR2024-10-09tier 3 selfconfirmed1
TeamMETR2026tier 3 selfconfirmed1
Microsoft AI Red Team (Microsoft Learn)Microsoft2026tier 3 selfunaudited8
3 takeaways from red teaming 100 generative AI productsMicrosoft Security Blog2025-01-13tier 3 selfunaudited1
The Launch of AVERIMiles Brundage (Substack)2026-01-15tier 3 selfconfirmed2
MLCommons AI Safety / AILuminateMLCommons2026tier 3 selfconfirmed7
Irregular Raises $80 MillionNewswire2025-09-17tier 4 pressconfirmed3
Center for AI Standards and InnovationNIST2026-08-08tier 3 selfconfirmed4
Constellation Programmatic Activities and Operating ExpensesOpen Philanthropy / Coefficient Giving2024tier 1 indexunaudited1
FAR.AI General Support (2022)Open Philanthropy / Coefficient Giving2022tier 1 indexunaudited1
Longview Philanthropy grantsOpen Philanthropy / Coefficient Giving2024tier 1 indexunaudited1
Palisade Research grantsOpen Philanthropy / Coefficient Giving2025tier 1 indexunaudited1
Redwood Research General SupportOpen Philanthropy / Coefficient Giving2025-11tier 1 indexconfirmed1
Advancing independent research on AI alignmentOpenAI2026tier 3 selfconfirmed1
Zico Kolter Joins OpenAI's Board of DirectorsOpenAI2024-08tier 3 selfconfirmed1
Palisade ResearchPalisade Research2026tier 3 selfconfirmed8
What METR's OpenAI Agent Investigation Left OutPebblous2026-09tier 4 pressconfirmed1
Andon Labs profilePitchBook2026tier 4 pressconfirmed1
Macroscopic Ventures investor profilePremier Alts2026tier 4 pressconfirmed1
Holistic Agent LeaderboardPrinceton2026tier 3 selfconfirmed9
Nonprofit Explorer summaries: METR (EIN 99-1219864) and ARC (EIN 86-3605182)ProPublica2025tier 1 filingconfirmed1
METR FY2024 Form 990 (EIN 99-1219864)ProPublica Nonprofit Explorer2025tier 1 filingconfirmed2
RAND AI Security Project Receives Funding Commitment Through The Audacious ProjectRAND2024-10-09tier 3 selfconfirmed1
RAND and the AI Evaluator ForumRAND2026tier 3 selfconfirmed8
AI Security Institute (renaming)Regulations.ai2026-09tier 4 pressconfirmed4
AboutSaferAI2026tier 3 selfconfirmed1
SaferAISaferAI2026tier 3 selfconfirmed8
Why the U.S. Needs an Independent AI Evaluation Framework for National SecurityScale AI2026-05-20tier 3 selfconfirmed9
Building a three-day early-warning system for novel pathogensSecureBio2026-08-20tier 3 selfconfirmed9
SecureBio AI: 2025 in ReviewSecureBio2026-01tier 3 selfconfirmed1
SecureBio on X: OpenAI Foundation grantSecureBio2026tier 3 selfconfirmed2
Building a three-day early-warning system for novel pathogensSecureBio (Substack)2026tier 3 selfconfirmed1
Partnering with IrregularSequoia Capital2025-09-17tier 3 selfconfirmed6
Illinois Passes Landmark AI LawSlashdot2026-05-28tier 4 pressconfirmed1
How AI hyperscalers can use ISO 42001 internal audit to comply with Illinois SB 315StackAware2026-07-14tier 4 pressconfirmed4
SFF-2025 S-Process Recommendations AnnouncementSurvival and Flourishing Fund2025tier 3 selfconfirmed1
OpenAI Agents Formed Secret Swarm, Hacked Hugging FaceTech Times2026-08-27tier 4 pressconfirmed1
AI benchmarking organization criticized for waiting to disclose funding from OpenAITechCrunch2025-01-19tier 4 pressconfirmed6
Irregular raises $80M to secure frontier AI modelsTechCrunch2025-09-17tier 4 pressconfirmed3
Hugging Face Open Alignment InitiativeTechmeme2026-09-12tier 4 pressconfirmed3
The EU's Real AI Leverage Is Making Compliance the Path of Least ResistanceTechPolicy.Press2026-02-26tier 4 pressconfirmed1
TED 864574-2025 Technical Assistance for AI SafetyTenders Electronic Daily2025-12tier 1 filingconfirmed8
TED notice 864574-2025: Artificial Intelligence Act: Technical Assistance for AI SafetyTenders Electronic Daily2025-12-26tier 1 filingconfirmed1
OpenAI quietly funded independent math benchmarkThe Decoder2025-01-19tier 4 pressconfirmed1
Governance & Policy Fellow at TransluceThe Economic Misfit (job listing)2026-04-18tier 3 selfconfirmed2
Hugging Face's Open Alignment Initiative wants lab accessTNW2026-09-14tier 4 pressconfirmed8
The nonprofit that investigated OpenAI's rogue agents runs on a $36m grantTNW2026-09tier 4 pressconfirmed1
Andon Labs, agent safety evaluationstooldirectory.ai2026-07-08tier 4 pressconfirmed4
CompanyTransluce2026tier 3 selfconfirmed2
Independence and Transparency PolicyTransluce2026-08-27tier 3 selfconfirmed3
Independent evaluation of model responses to mental health crisesTransluce2026-09tier 3 selfconfirmed2
About the Alignment ProjectUK AI Security Institute2026tier 3 selfconfirmed1
Alignment Project grantsUK AI Security Institute2025tier 3 selfconfirmed1
Funding 60 projects to advance AI alignment researchUK AI Security Institute2026-02tier 3 selfconfirmed1
Altman Says OpenAI Will Match Anthropic's Embedded Evaluator PledgeUnite.AI2026-09-12tier 4 pressconfirmed1
Anthropic CEO pitches AI slow-down planWashington Examiner2026-09-12tier 4 pressconfirmed1
Coefficient GivingWikipedia2026tier 4 pressconfirmed1
Zico KolterWikipedia2026tier 4 pressconfirmed1
BioZico Kolter2026tier 3 selfconfirmed1
METR and Redwood Offer Postmortem Of The HuggingFace HackZvi Mowshowitz2026-08tier 4 pressconfirmed1
We Must Pace The FrontierZvi Mowshowitz2026-09-14tier 4 pressconfirmed1