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. 20-GEMMASCOPE-RES-65K
    6. 780
    Prev
    Next
    INDEX
    Explanations

    closing syntax elements in code

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

    code blocks that denote the end of a process or function, specifically looking for closing statements and corresponding return outputs

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

    code punctuation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    CloseOperation
    -0.92
    __':
    
    -0.76
    WithIOException
    -0.75
    Autoritní
    -0.74
     متعلقه
    -0.71
    []){
    -0.69
    ")){
    
    -0.69
    __":
    
    -0.68
     **/
    
    -0.67
    '])){
    
    -0.67
    POSITIVE LOGITS
    ($"
    0.57
     cref
    0.56
     cours
    0.52
    altham
    0.51
    rawDesc
    0.50
    ($"{
    0.49
     base
    0.45
     hela
    0.44
    Cor
    0.44
     ویکی‌پدیای
    0.44
    Activations Density 0.005%

    No Known Activations