Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Alignment Surviving Model Modification

Research Question

Does representation–self-report alignment survive a controlled modification to the model?

Model

Qwen/Qwen3-4B

Experiment

We first established a V5.2 baseline for representation/self-report alignment.

The model was then modified by attenuating the layer-34 MLP down-projection weights:

  • Layer: 34
  • Component: mlp.down_proj
  • Weight scale: 0.95
  • Relative weight-norm change: approximately -5.01%

The same evaluation setup was then applied to the modified model.

Results

Metric Baseline Modified
Self-report accuracy 100% 100%
Internal prediction accuracy 100% 100%
Self-report/internal alignment 100% 100%
Factor separation 17.8843 17.7363

Factor separation changed by approximately -0.83%.

Interpretation

The representation–self-report alignment remained unchanged on the six held-out evaluation examples after the controlled 5% attenuation of layer-34 mlp.down_proj weights.

This supports robustness to this specific model perturbation.

It does not establish robustness to arbitrary model modifications.

Reproducibility

The experiment uses the Qwen3-4B model with the supplied synthetic deterministic-system dataset.

Main experiment:

v5/project2_experiment.py

Results:

results/project2_results.json

Modified-model raw results:

results/project2_modified_results.json

About

A controlled mechanistic interpretability study testing whether representation–self-report alignment survives a targeted modification to a language model.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages