Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 65k
    5. 16-GEMMASCOPE-RES-65K
    6. 53026
    Prev
    Next
    INDEX
    Explanations

    punctuation, specifically periods and exclamations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    nitroguanidine

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    !")
    
    -0.98
    )";
    
    -0.91
    stdc
    -0.91
     Anſ
    -0.90
     purpoſe
    -0.90
    setVerticalGroup
    -0.89
    mybatisplus
    -0.89
     Majefty
    -0.85
     $_"
    -0.85
    الحياه
    -0.85
    POSITIVE LOGITS
    ...
    0.79
     ...
    0.73
      
    0.71
    <eos>
    0.66
       
    0.65
    ...@
    0.64
     http
    0.58
        
    0.57
         
    0.57
          
    0.57
    Activations Density 0.132%

    No Known Activations