seanhagstrom commited on
Commit
ae0dcb0
·
verified ·
1 Parent(s): 108b4e9

presence: sync card

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -81,7 +81,7 @@ Instead of retraining or repeating fine-tuning, Authentrics works directly with
81
 
82
  **Demonstrates:** catastrophic forgetting, regression recovery, model optimization
83
 
84
- - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-model-analysis-experiments)
85
  - [Model on Hugging Face](https://huggingface.co/authentrics/ztom-gemma3-1b-recovery)
86
 
87
  ---
@@ -96,7 +96,7 @@ Authentrics identifies where the unwanted learning was introduced and removes it
96
 
97
  **Demonstrates:** memorization analysis, training attribution, machine unlearning
98
 
99
- - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-analysis-examples)
100
  - [Model on Hugging Face](https://huggingface.co/authentrics/pii-removal-medical-llama32)
101
 
102
  ---
@@ -113,7 +113,7 @@ Authentrics analyzes the model's internal structure and behavior to expose routi
113
 
114
  **Demonstrates:** model interpretability, MoE routing, internal model analysis
115
 
116
- - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-model-analysis-experiments)
117
  - [Model on Hugging Face](https://huggingface.co/authentrics/mixtral-expert-routing)
118
 
119
  #### Find Where a Model Changed Across Its Training Run
 
81
 
82
  **Demonstrates:** catastrophic forgetting, regression recovery, model optimization
83
 
84
+ - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-analysis-examples/blob/main/hf_legal_understanding.ipynb)
85
  - [Model on Hugging Face](https://huggingface.co/authentrics/ztom-gemma3-1b-recovery)
86
 
87
  ---
 
96
 
97
  **Demonstrates:** memorization analysis, training attribution, machine unlearning
98
 
99
+ - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-analysis-examples/blob/main/hf_medical_chatbot.ipynb)
100
  - [Model on Hugging Face](https://huggingface.co/authentrics/pii-removal-medical-llama32)
101
 
102
  ---
 
113
 
114
  **Demonstrates:** model interpretability, MoE routing, internal model analysis
115
 
116
+ - [Source code (GitHub)](https://github.com/Authentrics-ai/authentrics-model-analysis-experiments#mistralmixtral-moe-expert-activation)
117
  - [Model on Hugging Face](https://huggingface.co/authentrics/mixtral-expert-routing)
118
 
119
  #### Find Where a Model Changed Across Its Training Run