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-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 56698
    Prev
    Next
    INDEX
    Explanations

    patterns in function calls and variable assignments

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

    X'th term or roller

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     ProtoMessage
    -0.47
     Vikipedi
    -0.45
    Читати
    -0.44
    zarchiwizowane
    -0.40
    épar
    -0.39
    PMailer
    -0.37
     regalías
    -0.37
    nalités
    -0.36
     meilleures
    -0.36
    université
    -0.36
    POSITIVE LOGITS
    setVerticalGroup
    0.78
     enumerate
    0.70
    addPreferredGap
    0.67
    didSet
    0.58
    enderror
    0.55
    writeField
    0.51
    ViewFeatures
    0.49
     thumbs
    0.48
    onViewCreated
    0.48
    complexContent
    0.47
    Activations Density 0.004%

    No Known Activations