wilson-interval

A comprehensive module used to calculate the high bound, low bound, and center of a Wilson score interval.

Synthetic, dimensionless testbench for the signed-amplitude sine encoding in M. S. Lee's 'Digital to Wave' preprint: seeded noise/phase/CFO/timing benchmarks, Wilson CIs, hash-locked modules. No physical validity claimed.

能力衰变哨兵 — baseline per-capability success from early turns and alert when a class decays mid-run (is_error + Wilson, fully offline). v0.2.0 adds the guard CI gatekeeper: 0/1/2 exit codes + GitHub Actions annotations, plus clean bad-input error handling.

various confidence interval implemented in javascript

Generated code, actually checked. A measurement harness for a code-generation agent: typecheck and bundle signals, Wilson intervals, paired comparisons.

Brings no notion of a correct answer and wants none: a probe asks whether your own validator would have accepted the response, counts the attempts the shipped retry loop really spent, and reports a tie when two configurations cannot be separated.

Synthetic engineering-validity study of the THz human-wave interface equations in Digital to Wave Part II: equation inventory, dimensional checks, protocol-locked channel benchmark. No physical claim.

Wilson confidence intervals, difference of proportions, and McNemar's test for LLM evaluation results. Client-side, single file.

Read a landing page or fake door test against the success line written before launch: passes, fails or undecided, with Wilson ranges, variant comparisons and the visitors still needed. No dependencies.

Domain-pluggable engine for competitive-play analysis — games plug in via a six-point contract: percentiles, Wilson intervals, meta drift. Adapters: Dota 2, chess.

Prove a flaky-test fix: run base and head paired, many times, and get FIXED / NOT FIXED with a p-value

Classical toy simulation (Stim + PyMatching, classical CPU only) of surface-code and repetition-code memory scaling. No hardware, not a QEC demonstration.