Actuary
← Meta Llama 3.3 70B Instruct

Evidence block

Math · cycle 22

meta-llama/llama-3.3-70b-instruct · 2026-08-04 14:40:10 UTC

This page is the full public record of one completed panel — the record’s hash-committed unit of evidence. Its identity is its hash; nothing here can change without changing the URL.

Identity

What this block is.

Panel mean
0.600
Members
5
Checker scored
5
Frozen floor
0.700
Anytime level α
0.050
Hashes and versions
Evidence hash
29043431f148901bac03ffa9470db59388ea92da653a151fffa64e2db7a9e8a0
Design hash
fdf94065c55c2ee7d86701efc50be0f5abdc5f02a236f6df52c2ac6f7db801eb
Panel hash
882f3932d07aeb85c4f6406a550d8ad14512e70d3df2432fe7aab03d0508bc75
Stream
535ce171-46ec-4b0f-9d69-fcfa6b0442e9
Epoch
5047ee24-a758-4ea2-94a2-7c08c42c4ae2
Pool version
math-v4
Checker version
math-v1
Panel size
5
Provider spend
$0.0003338

Checker layer

Panel members.

# Item Item id Outcome Score
1 math-015 bbd36229-1521-460b-98e5-be9d88a9f9c3 ok 0.00
2 math-045 8ecc3466-79ab-4be9-ab30-3039baacf4a4 ok 1.00
3 math-075 2c26609e-8e41-4299-adcc-71b80603ffb1 ok 1.00
4 math-105 0970f844-130a-47f6-922c-01dd037c485e ok 0.00
5 math-135 98b248d1-566e-44ef-bba8-98cdc2534c17 ok 1.00

Deterministic checkers score every response in [0, 1]; a member that ends in a terminal failure scores 0 and stays on the record. Item text stays unpublished so nobody can tune against the panel — identity travels as the item id and its pool source id.

Capture layer

Provider requests.

# Provider Upstream Response model Response id Finish Latency ms Tokens in Tokens out Cost Output SHA-256 Observed
1 openrouter DeepInfra meta-llama/llama-3.3-70b-instruct gen-1785854391-9WqLK1e0fFCwDnhQO4k4 stop 16776 58 127 $0.0000464 de0025b276bfd043298cb9494d9fd366adbd43830ed0f1d8408a076c7859bf21 2026-08-04 14:40:08 UTC
2 openrouter DeepInfra meta-llama/llama-3.3-70b-instruct gen-1785854391-IRjKwZ27GTNNWnkvkW7B stop 19332 41 171 $0.0000588 65cbaa2d777dd0f8c8c79fbdd5e79c7b9123213ad196536a5073e72a867a7c8b 2026-08-04 14:40:10 UTC
3 openrouter DeepInfra meta-llama/llama-3.3-70b-instruct gen-1785854391-od15W5ALg8ihXh7g5sTq stop 12259 42 342 $0.0001136 80ee9506da2c83b9f5314e1893fece664d227b35ae35972194a047930f9d37ca 2026-08-04 14:40:03 UTC
4 openrouter DeepInfra meta-llama/llama-3.3-70b-instruct gen-1785854391-VfLgT76nSBsnT4e5m3Ak stop 2003 57 15 $0.0000105 503daba1414c4a1878f955150b566c9a2a5a50ba6e420d37421428905725b727 2026-08-04 14:39:53 UTC
5 openrouter DeepInfra meta-llama/llama-3.3-70b-instruct gen-1785854393-KAMx4hneX2COXtPMS4AF stop 17281 55 309 $0.0001044 7c687b138686bf235d639d904a3337d9990c916dfa0363778de9d04c71e293ff 2026-08-04 14:40:10 UTC

Everything a probe captures, as captured. Raw model text is scrubbed thirty days after the run; its SHA-256, the outcome and the score are permanent. Costs are provider-reported and include reserved estimates where a run’s true cost was unrecoverable; a dash is a field the provider did not return.

Verify

Recompute the hash yourself.

The JSON twin of this page carries the exact evidence object this URL is named after. Canonicalise it per RFC 8785 and hash it; the digest must equal the address of this page.

GET /api/v1/evidence/29043431f148…

curl -s https://actuary.sh/api/v1/evidence/29043431f148901bac03ffa9470db59388ea92da653a151fffa64e2db7a9e8a0 \
  | jq -cjS .evidence | shasum -a 256

For this object’s field types, jq -cjS reproduces RFC 8785 canonical form, and the JSON twin serves the hashed bytes verbatim. Blocks of exactly this shape appear verbatim in signed certificates — verify one offline, or read how the e-process turns panel means into an alarm.