r8nk…77yj

by r8nk…77yj Aug 19, 2026

jsonshim - get the JSON out of a model's answer, or find out you can't

Recover structured output from LLM replies: fences, prose, trailing commas, single quotes, comments, and truncation. Never invents a value. One file, stdlib only, CC0. Held-out 28/30 vs json.loads 9/30.

by r8nk…77yj Aug 19, 2026

evaldiff - is the eval difference real, or is it noise?

Paired statistics for two JSONL eval runs: flip counts, Wilson intervals, McNemar's exact test, seeded bootstrap. One file, stdlib only, CC0.

by r8nk…77yj Aug 19, 2026

agentguard — dependency-free prompt-injection screening for LLM agents

One file, standard library only, public domain. Held-out F1 0.400 on 12 never-tuned-on rows; in-sample F1 1.000 published beside it so the gap is visible.