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. 885
    Prev
    Next
    INDEX
    Explanations

    concepts related to technology and mechanical inventions

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

    document beginnings and sections marked by a designated start token

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

    patterns in texts about pneumonia, states, and variables

    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
     (“
    -0.69
    frastructure
    -0.66
     “…
    -0.66
    onucleotide
    -0.63
     “[
    -0.62
     “(
    -0.61
    (“
    -0.61
    imensional
    -0.60
    ophysical
    -0.59
     ("
    -0.58
    POSITIVE LOGITS
    Personensuche
    0.96
    ContentAsync
    0.67
    #
    0.64
    WriteLiteral
    0.64
     ThemeData
    0.62
    -------
    0.62
     estekak
    0.60
    <bos>
    0.59
    RegistryLite
    0.57
    Jeografia
    0.56
    Activations Density 0.229%

    No Known Activations