recursiveauto
/

Symbolic-Residue-Interpretability-Powered-By-Failure-Not-Completion

interpretability

constitutional AI

transformer-failure-analysis

refusal-diagnostic

Model card Files Files and versions

recursiveauto commited on Apr 6

Commit

d26250d

·

verified ·

1 Parent(s): d4556d9

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -13,7 +13,7 @@ tags:
 <div align="center">
   # On the Symbolic Residue of Large Language Models:
-  # Diagnosing and Modeling Biological Failure Traces in Local Replacement Models
 ## **─ What If Interpretation Itself is Biased By Internal Salience and Conflict Resolution? ─**

 <div align="center">
   # On the Symbolic Residue of Large Language Models:
+  # The Missing Biological Knockout Experiments in Advanced Transformer Models
 ## **─ What If Interpretation Itself is Biased By Internal Salience and Conflict Resolution? ─**