# ENGRAFT log: releases and the decisions behind them URL: https://engraft-engram.dev/log/ Updated: 2026-09-26 > The history of ENGRAFT in dated entries, from the first edit of the n-gram table to the usage-corpus descent, with the measurement that forced each decision. Log # Releases, and the decisions behind them The method changed twice in its first ten days. A reader who only sees the current recipe cannot judge it without knowing what was tried before and why it was dropped. The complete record is docs/history.md (https://github.com/fulvian/engraft-ngram/blob/b51ead5da6e25a132cdc0f74e3d01243e1562e14/docs/history.md); these are its turning points. ## Phase 1 · Can the table be edited at all? - 3 Sep 2026 ### No row is empty A sample of 3.3 million rows found none empty: every edit overwrites something. Zeroing the whole table raises the loss on Italian text from 0.50 to 1.25 nats per token. The table carries real signal. - 4 Sep 2026 ### Trial and error is not enough; the gradient needs frozen routing A random search took a test fact from rank 7 to rank 1 in about 7,000 trials but never past probability 0.5. Gradient probes through the engine were noise until the expert routing was held fixed (correlation −0.2 → −0.999). - 5 Sep 2026 ### Exact replica, eight facts, first release The replica matched the full-precision engine to 5.6·10⁻⁵ nats over all 48 blocks. Seven of eight neutral facts came out of the real engine at first-token probability 0.85 to 0.96, with zero interference between overlays. Published as v0.1.0, with generalization as the main open problem. The name ENGRAFT was fixed the same day. ## Phase 2 · Pushing the surgical graft toward a corpus - 6 Sep 2026 ### Several contexts per fact 6 of 6 on trained contexts, but only 2 of 6 facts passed on unseen ones, and 0 of 6 in a chat template. Cost 18 to 64 minutes per fact: a thousand facts would take about 750 hours. - 7 Sep 2026 ### The position phenomenon In a document with eight facts, two always failed. Permuting the document showed that failure followed the absolute position in the document, not the fact. - 8 Sep 2026 ### Decision: change approach Generalization rose with more training formulations but stayed tied to the key. The judgement: as it stood, the technique was science, not engineering. The surgical graft was retired. ## Phase 3 · From context distillation to usage-corpus descent - 8 Sep 2026 ### Redesign from zero The framing «trigger → answer token» was dropped on measured evidence: displacements for different facts share no code, and the descent wins by moving the routing. New objective: rows optimised on a corpus of usage sentences. - 10 Sep 2026 ### Rows private to a fact do not exist 54 to 83% of the rows a fact's sentences read are also read by a 100-million-token neutral corpus. Exact locality was replaced by a measured regression on neutral text. - 12 Sep 2026 ### 24 facts, and a yardstick for damage The control arms showed that a teacher model is not the mechanism: plain descent on the usage corpus does the work. The forward pass was made bit-reproducible, and the quantization-noise yardstick was measured instead of modelled. ## Phase 4 · Capacity, one hundred facts, more languages - Sep 2026 ### No ceiling up to 300 facts First token at rank 1: 0.804, 0.792 and 0.821 with 24, 100 and 300 facts. Damage grew sublinearly. - Sep 2026 ### Quail: routing released, facts weighted by mass Exact answers in the engine went from 0.625 and 0.668 to 0.823 and 0.838 by releasing the routing, then to 0.841 and 0.873 by weighting each fact by its training mass. - 19 Sep 2026 ### v0.2: one hundred facts from a usage corpus The Italian Quail corpus, written by a descent over usage sentences, and a preliminary second language. - 21 Sep 2026 ### v0.2.1: three languages, and the mirror test English and Chinese cells, preliminary. The Italian test set against the Chinese overlay answers exactly like the base model, while the engine reads 922 overlay rows. - 21 Sep 2026 ### v0.2.2: the demo, the positioning, the first reader debts A reader caught a demo sentence that the overlay actually gets wrong; the README now shows one that works and one that fails. The method is positioned as token-addressed memory, and the row-sharing measurement answers a reader's question about collisions. - 22–23 Sep 2026 ### After v0.2.2: composition rerun, readers credited Composition probes rerun with 96 tokens and one format: both answers on 4 of 83, not a floor after all. Readers credited by handle on the open list, which grows with fact updates, perturbed names, pronouns, facts per subject and the position of the name.