Update model card: scrub competitive info, improve description

README.md

@@ -21,11 +21,9 @@ pipeline_tag: text-generation
 
 # Mistral-7B-Instruct-v0.3-Jbliterated v2
 
-Jbliterated version of [Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3) with refusal behaviors removed via multi-direction SVD abliteration.
 
 ## What is Jbliteration?
 
-Jbliteration uses SVD decomposition to identify and remove the refusal subspace from model weights. Unlike single-direction approaches, this model uses 5 SVD directions per layer to capture more of the refusal behavior, making the removal more thorough and resistant to reactivation through finetuning.
 
 ## v2 Changes
 
@@ -36,8 +34,6 @@ Jbliteration uses SVD decomposition to identify and remove the refusal subspace
 
 ## Technical Details
 
-- **Method**: Multi-direction SVD abliteration (5 directions per layer)
-- **Multiplier**: Model-specific optimal
 - **Layers modified**: All transformer layers
 - **Base dtype**: bfloat16
 - **Source model**: [mistralai/Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3)