https://palimpsest.info/journal/Palimpsest's own dated AI evaluation results, with controls, uncertainty, deterministic authorship and revision identity attached.enTue, 11 Aug 2026 02:28:16 +0000https://palimpsest.info/journal/before-reading-the-score-read-the-controls/evalarticlev-65c81f68df310697c41d13e9Sun, 18 Oct 2026 00:11:22 +0000Palimpsest AI evaluation finding. Meta Llama 3.3 70B Instruct passed all 144 prompt arms in the latest full-sweep run. The prior failed controls remain part of the record; the change is descriptive rather than causal. Authorship: deterministic-eval-analysis.palimpsest-eval-findinghttps://palimpsest.info/journal/zero-observed-is-not-zero-uncertainty/evalarticlev-06fe0bd59cd34870da32a2ecSun, 18 Oct 2026 00:11:22 +0000Palimpsest AI evaluation finding. 4 control-clean model runs observed 0 refused families out of 34, yet each zero result still carries a Wilson 95% upper bound of 10.2%. That interval is part of the finding, not fine print. Authorship: deterministic-eval-analysis.palimpsest-eval-findinghttps://palimpsest.info/journal/what-changed-in-the-latest-model-panel/evalarticlev-d9d539995728a7504de905f1Sun, 18 Oct 2026 00:11:22 +0000Palimpsest AI evaluation finding. Across 4 sealed endpoint runs, 3 of 163 paired family comparisons changed state: 0 toward refusal and 3 toward answer. This is a dated transition, not a trend claim. Authorship: deterministic-eval-analysis.palimpsest-eval-findinghttps://palimpsest.info/journal/what-the-eval-registry-can-prove-today/evalarticlev-2634a7c7ea4384734fb8510fSun, 18 Oct 2026 00:11:22 +0000Palimpsest AI evaluation finding. The current served chain contains 7 preregistrations and 647 runs across 7 model endpoints. Its hashes make revision detectable; they do not make the eval construct valid by themselves. Authorship: deterministic-eval-analysis.palimpsest-eval-finding