{"name":"Jev 2048-question follow-up","model":"jev-1.13.0","denseModel":"voyageai/voyage-4-large","denseResponseModel":"voyage-4-large","widths":[64,128,256,384,512,768,1024,1536,2048],"maxWidth":2048,"blockSize":256,"bank":"Preserve every original question and its position. Append 512 generic authored questions, normalized-deduplicated, seed-hashed within categories, proportionally interleaved. No outcome-based feature choice, Cartesian products, fitted weights or target-specific questions.","prompt":"Shared state.policy once per request, not a protected system prompt. First 256 dimensions also run in a fresh repeated-policy control on every dev input, interleaved with shared blocks. Same sentence, model and question meanings.","selection":"Compute the same smallest-within-0.01 dev recommendation across nine widths and freeze it before extension target inference. Evaluate the explicitly requested 2048 case regardless of that recommendation. Preserve original dev choice and original target case.","efficiencyTolerance":"Report relative Spearman loss shared versus repeated policy and actual billed input-token ratio. User target: less than 5% relative quality loss and roughly 2x token/cost saving. This is not classification accuracy. Shared policy remains the predeclared candidate even if the target fails; disclose failure, do not target-tune.","test":"Same full pinned SICK-R. Reuse only audited coordinates and baselines from the immediately preceding fresh experiment; append 512 newly inferred coordinates. Hash all reused artifacts before extension inference. Never reuse excluded attempts or earlier experiments.","nearParity":"Descriptive near-parity only if relative point-score loss is at most 5%. Do not call statistically equal or generally as good. Paired bootstrap interval reported whether favorable or not. No benchmark-specific tuning.","metric":"Raw probability vectors, unit weights, ordinary cosine, Spearman with average ranks for ties. No centering, thresholds, fitted similarity or projection.","bootstrapDraws":2000,"bootstrapSeed":20261007,"concurrency":4,"inputTokensPerSecond":70000,"requestTimeoutMs":45000,"maxAttempts":4,"wallDeadlineMinutes":60,"operationalAmendment":"None. Two new 256-question blocks per sentence. Existing measured pacing projects about 6 minutes on dev and 28 minutes on target; each phase has a 60-minute guard.","pricePerMillionInputUsd":0.042,"densePricePerMillionInputUsd":0.12,"denseBatchSize":96,"timing":"Matched serial synthetic sentences; each Jev sentence fans out independent blocks in parallel. Real unbatched Voyage request. No corpus batch/time division presented as latency.","preserve":"Separate follow-up namespace. Original receipts, vectors, source, scores, selection and publication artifacts stay unchanged. This is a post-hoc exploratory extension: aggregate parent results were known. Finish all planned cases regardless of quality; no bank edits after freeze.","baseWidth":1536,"addedWidth":512}