Promythicus (@promythicus) profile picture

AI research lab and maker of private AI servers for professional firms in Canada. Your documents never leave your office. We publish what we test.

Canada
Joined November 2026
A local LLM now answers contract questions with sentence numbers, not typed quotes. Clauses found: 82% -> 86%, twice as fast, every quote exact. Joined with a sheet the model writes at filing: 87%. RAG: 50%. Built on work by @HyunjongOk @jaeho_lee_. github.com/Promythicus/joine…
1
1
1
11
We edited 24 facts into a model's weights. Change the system prompt and recall fell from 90%+ to 26%. Then the model wrote them as a table instead: 100% in every setting, no training. Built on work by @HyunjongOk @jaeho_lee_ and @y_mehtu @csinva. github.com/Promythicus/writt…
1
1
1
8
Chat, what do you think about private inference servers?
3
We wrote 24 facts straight into a model's weights with one math step. It recalled over 90% of them. Then we changed the system prompt and it recalled 26%. Setup: made-up facts about a fake company, one layer of a quantized 27B (a compressed model that runs on a normal GPU).
1
1
1
2
97
We applied weight-space programming, fixed a broken model by editing 128 of its neurons by hand. 9% -> 39% on math problems, no training. editing weights directly. The model (Qwen3.5-0.8B) gets stuck "thinking" forever. We redid a published fix without looking at it.
1
1
1
15
An LLM's own hidden states beat a dedicated embedding model at search. Nothing trained. Hidden states = the numbers inside the model while it reads. We averaged them over each passage, at a middle layer of a 27B model, and used them to find the passages that answer a question.
1
1
1
64
We took the embedding model out of our RAG pipeline and accuracy went up 10 points. In its place: when a document comes in, the LLM writes one line per chunk saying what that part covers. For a question, it reads those lines and picks which chunks to open. No training.
1
1
1
20
Just putting the whole document in the prompt beat RAG by 11 points on our test. RAG = search for the relevant passages first, read only those. We ran both on a local 27B model, over contracts where lawyers had marked the answers (CUAD). Whole document: 82% found. RAG: 71%.
1
1
1
88