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. 812
    Prev
    Next
    INDEX
    Explanations

    the beginning of new sections or topics

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

    occurrences of printing actions or commands in programming contexts

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

    aller voir

    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
     싶
    -0.51
    لال
    -0.51
     code
    -0.48
    طل
    -0.47
    ној
    -0.47
     kasa
    -0.46
    zz
    -0.44
     line
    -0.44
    soort
    -0.44
    ьаж
    -0.44
    POSITIVE LOGITS
     AspNetCore
    0.76
     surla
    0.71
     uintptr
    0.67
    ruptedException
    0.65
     الحره
    0.64
    ViewChild
    0.61
    complexType
    0.60
    ratulations
    0.59
    idł
    0.57
    ChildScrollView
    0.56
    Activations Density 0.011%

    No Known Activations