DietrichGebertandClaude Opus 4.8 2e6a93765a fix(benchmarks): count unfenced code, ASCII-safe output, refresh llama3.2 results (#67) ...

Fixes the local benchmark LOC counter (counted only fenced code, scored bare output 0), makes summary output ASCII-safe (a Unicode arrow crashed the script on Windows cp1252), gitignores generated artifacts, and refreshes the llama3.2 writeup with n=5 data showing the LOC effect is within the noise floor. Follow-up to #63. Verified live.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-15 16:22:32 +02:00

..

2026-06-12-caveman-vs-ponytail.md

feat: skill v3 — compress SKILL.md 115 to 95 lines

2026-06-12 03:58:44 +02:00

2026-06-12-v4-hardening-vs-caveman.md

docs: same-model control arm, refresh numbers and chart

2026-06-12 12:44:30 +02:00

2026-06-15-llama3.2-local.md

fix(benchmarks): count unfenced code, ASCII-safe output, refresh llama3.2 results (#67)

2026-06-15 16:22:32 +02:00