3 Commits

Include renames

Author	SHA1	Message	Date
rzen	921069c5db	Add degradation analysis + backing artifacts ... - reports/degradation-analysis.md: interpretation of the one-parameter-at-a-time ablations (ctx/short/small/data vs v1), grounded in val loss + sample text. Key findings: held-out loss tracks quality for generalizing models; different degradations give qualitatively different failure text; data-starvation overfits (train ppl 1.1 / val ppl 322) with samples that hide the damage. - reports/compare.md: side-by-side samples across all configs - reports/loss-{small,short,ctx,data}.csv: variant training curves	2026-07-12 18:28:43 -04:00
rzen	84c41bc612	Phase 5: v1 sample gallery — iteration 1 complete ... - reports/samples-v1.md: 5 story-opening prompts sampled at temp 0.8 / top-k 200. Grammatical, coherent story arcs; expected sub-sweet-spot wobbles (plot logic, entity drift). All iteration-1 definition-of-done items now checked off.	2026-07-12 12:44:10 -04:00
rzen	ba73c8a7e8	Phase 4: v1 training run results (val loss 1.52, ppl ~4.6) ... - reports/loss-v1.csv: full train/val loss log over 20k steps - reports/loss-v1.txt: ASCII loss curve (steep descent then flattening tail) Training ran ~58 min on M1 MPS at 5.7 it/s. No overfitting (train~=val).	2026-07-12 12:35:47 -04:00