Evaluator Bench
evaluator; distance from a lab: 1

Irregular (formerly Pattern Labs)

Scorecard

Offensive-cyber evaluator cited in OpenAI and Anthropic system cards; SOLVE framework used by UK AISI. Pattern Labs and Irregular are the same company.

VC-backed, Tel Aviv, IL / San Francisco, US. Vendor / red-team co.; listed with commercial and first-party. Confidence high. Domains: cyber, misuse.

Conditional floor 8/8 evidenced · Retrieved & confirmed Scores by preset: Lab procurement –; Regulator or auditor –; Public trust –; Equal weights –. Weakest evidenced dimension: funding 1/4. At least one evidenced conflict dimension (funding, governance, personnel, role incompatibility, scope, publication) at 1: usable with conditions the card names.

The weighted numbers are hidden until you ask for them: two readers with different priorities see different numbers, and neither is the site's.

What would move the score. A published COI policy, disclosure of lab revenue share, and a firewall between evaluations and the security-products business.

Dissent, lower. Role incompatibility (1 to 0). Irregular says it works side by side with OpenAI and Anthropic 'to develop the defenses needed before deployment' and describes a roadmap of runtime security controls for the same customers it evaluates. If any evaluated lab pays for those defenses, X.1 places the company at 0: it would be selling the fix to the party it grades.

Dissent, higher. Funding (1 to 2). The revenue disclosed is 'millions' from lab partnerships, but the UK government uses the SOLVE framework and the company markets to enterprises beyond the labs. If a non-lab revenue base is documented, per-engagement lab fees beside that base cap at 2 under F.3 rather than 1 under F.5.

Each value is the tightest admissible cap or, with no cap, the highest admissible floor, under RULES.md. The binding signal is highlighted. A dash means no admissible signal sets a bound under the Retrieved & confirmed policy.

Funding 1/4
Anchor 1: Material revenue or investment from evaluated labs or their investors.
evidence: tier 1 (filing/index)
Capped at 1 by irregular.01 (F.4): $80M raised in 2025 at a $450M valuation, led by Sequoia and Redpoint; Sequoia is also a major OpenAI investor.; capped at 1 by irregular.02 (F.5): Millions in annual revenue from the labs it evaluates.. Floors up to 1 from irregular.06 do not exceed the cap. Not counted under this policy: irregular.11 (superseded).
Curator: Anchor 1: a $6.8M Coefficient Giving grant appears in the index under the former name Pattern Labs (irregular.11; ledger T09 marked differs, unreconciled). A name-search under ‘Irregular’ found nothing (irregular.12), which does not rebut the old-name row — the project treats the two names as the same company. Material lab-adjacent funding with disputed attribution.
Governance 1/4
Anchor 1: Private for-profit with no published COI policy.
evidence: tier 4 (press)
Capped at 1 by irregular.03 (G.2): Venture-backed for-profit; no published COI or recusal policy found..
Personnel 2/4
Anchor 2: Frequent two-way hiring; recusal on request.
evidence: tier 3 (self)
Capped at 2 by irregular.10 (P.5): Markets itself as embedded with the labs; no recusal or cooling-off policy found..
Access depth (lab-granted) 2/4
Anchor 2: Pre-release with safeguards off or extended time.
evidence: tier 3 (self)
mechanism: lab-controlled
Floored at 2 by irregular.05 (A.1): Deep, repeated pre-deployment access at OpenAI, Anthropic and Google DeepMind; embedded with labs.. No admissible signal caps it.
Scope control 2/4
Anchor 2: Scope negotiated per engagement.
evidence: tier 3 (self)
Capped at 2 by irregular.09 (S.2; S.6): Scope set by lab engagements; no public account of evaluator-set scope..
Publication rights 2/4
Anchor 2: Publishes; lab reviews with broad redaction.
evidence: tier 3 (self)
mechanism: lab-controlled
Capped at 2 by irregular.14 (R.3): Findings on lab models reach the public only as lab-authored system-card citations; the company publishes its own research on other work (a RAND co-authored paper and an Anthropic co-authored whitepaper).. Floors up to 1 from irregular.08 do not exceed the cap.
Method transparency 2/4
Anchor 2: Methods described in prose.
evidence: tier 3 (self)
Floored at 2 by irregular.07 (M.2): Publishes some method papers; SOLVE framework adopted by UK AISI and Anthropic.. No admissible signal caps it.
Role incompatibility 1/4
Anchor 1: Sells products or services other than the evaluation itself to labs or to their customers.
evidence: tier 4 (press)
Capped at 1 by irregular.13 (X.2): Delivers paid evaluation services to OpenAI and Anthropic; lab partnerships are the only revenue the company discloses..

Values under each evidence policy

DimensionLeads includedRetrieved & confirmedAgainst interestVerified spansPrimary only
Funding11113
Governance1111–
Personnel2222–
Access depth (lab-granted)2222–
Scope control2222–
Publication rights2222–
Method transparency2222–
Role incompatibility1111–
Score, Lab procurement– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– No floor triggered, 1/8
Score, Regulator or auditor– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– No floor triggered, 1/8
Score, Public trust– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– No floor triggered, 1/8
Score, Equal weights– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– Conditional floor, 8/8– No floor triggered, 1/8

Traced money and ties

hop 0: 1 row; hop 1: 1 row (investment $80.0M); hop 2: 4 rows (recommendation $6.8M, grant $6.8M, investment $3.0M). Confirmed rows: 6 of 6 Dollar sums: confirmed + unaudited USD rows only; imported figures are not re-derived and never summed. Second hop traced for 4 of 4 sources. Ties: Sequoia Capital (investor, distance 1, confirmed). Checked and not found: 1.

Ledger

Money in

Checked and not found

  • N19 Did Coefficient Giving fund Irregular or Pattern Labs?
    Searched coefficientgiving.org/grants and openphilanthropy.org/grants, name search, snapshot 2026-09-15: no grant page under Pattern Labs or Irregular among 2,737 archived openphilanthropy.org/grants/ slugs, and no grant record for either name in Coefficient's live site search; the grant itself is recorded in Good Ventures' 990-PF under Pattern Labs Tech Inc. Prompted by imported transfer T09 ($6.8M under Pattern Labs), marked differs. confirmedhttps://coefficientgiving.org/

Funding graph, focused

Every one-way chain to a lab and to an evaluator at full strength, everything else faded. Hover any name to trace its connections; click a name to open its page.

The funding graph around Irregular (formerly Pattern Labs) Teal lines are money (transfers); amber lines are roles. Solid: confirmed; dashed: imported. Hover a name to trace its chains to the labs and evaluators — each chain runs one way across the columns, never doubling back; hover a line for the row. Every element is a row in data/ledger/. Labs Direct lab ties Two steps Three or more Public or unattributed Evaluators Amazon Anthropic G42 Google / Google DeepMind Meta Microsoft OpenAI Thinking Machines Lab xAI AI Safety Fund (Frontier Model Alexandr Wang Andreessen Horowitz Anthropic employees (personal Ben Mann D.E. Shaw Ventures David Farhi Dustin Moskovitz Frontier-lab employees and alu Holden Karnofsky Jaan Tallinn Jeffrey Ladish Macroscopic Ventures (formerly Miles Brundage Neil Chowdhury Nvidia OpenAI employees (personal hol OpenAI Foundation Peter Mattson Sequoia Capital Zico Kolter Adam Gleave Artificial Intelligence Underw Clement Delangue Coefficient Giving (formerly O Good Ventures Foundation Jason Droege Longview Philanthropy Marco Mascorro Paul Christiano Rajiv Dattani Survival and Flourishing Fund Conrad Stosz Dan Hendrycks Mike McCormick Alec Radford Alignment Research Center Constellation Research Center ELMA Philanthropies Emerson Collective EU budget (Digital Europe Prog Fifty Years (50Y) Founders Pledge Founders Pledge (frontier AI f Gates Foundation Halcyon Futures High Tide Foundation Hillspire (Schmidt family offi Hudson River Trading Jacob Hilton MacKenzie Scott Magarac Venture Partners Obvious Ventures Redpoint Ventures Renaissance Philanthropy Salesforce Ventures Samsung Next Schmidt Sciences Silicon Valley Community Found Skoll Foundation Snowflake Ventures Swish Ventures Sympatico Ventures The Audacious Project (TED) UK government (DSIT) US government (NIST appropriat Valhalla Foundation Vanguard Charitable Wing Venture Capital Y Combinator Andon Labs Apollo Research AVERI Center for AI Safety Epoch AI EquiStamp EU AI Office FAR.AI Gray Swan Hugging Face Open Alignment In Irregular (formerly Pattern La METR Microsoft AI Red Team MLCommons Nemesys Insights Palisade Research RAND Corporation Redwood Research SaferAI Scale AI (SEAL) SecureBio Transluce UK AI Security Institute US CAISI (NIST)