Skip to content

Model Mapper 0.1.0

Choose a tag to compare

@jphoenix-dev jphoenix-dev released this 19 Oct 10:05
· 1 commit to main since this release

Model Mapper 0.1.0: a clean, tested package. Give it a concept; it finds which neurons of a language model the
answer depends on, by switching neurons off one at a time against random neurons switched off the same way.

  • modelmapper check verifies an install against our numbers and writes a report you can send back.
  • modelmapper map-concept "X" / map-batch map one concept or a list (resumable); steer turns chosen neurons up;
    lookup reads published records. Python API: from modelmapper import map_concept.
  • Every record carries its provenance (GPU, software versions), and neurons near the top-30 cut-off are labelled
    uncertain using a band calibrated across GPU types (METHOD.md).
  • Tested on two machines: a Windows 10 desktop (CPU, float32) and a Windows 11 laptop (RTX 1000 Ada 6 GB, bfloat16,
    Python 3.10), both passing check --apis with the same top neurons (CALIBRATION.md).
  • FINDINGS.md: what we have found on Qwen3-8B, how sure we are, and the corrections made after an
    independent review (some figures are provisional while the last 1,000 concepts finish).

This replaces the research history that was in this repository. The full record of the 18 Oct 2026 release, every
script and result included, stays at doi:10.5281/zenodo.23459926. The
explainer and the Qwen Neuron Map now live at jphoenix-dev/eiffel-tower-qwen.