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. 40265
    Prev
    Next
    INDEX
    Explanations

    terms related to scientific and technical measurements

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

    scientific and technical terms

    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
    <h1>
    -0.30
    “
    -0.28
    「
    -0.27
    『
    -0.26
     
    -0.26
    #!/
    -0.25
    <strong>
    -0.25
    from
    -0.25
    <b>
    -0.25
    <?
    -0.24
    POSITIVE LOGITS
    principalColumn
    1.23
    <unused74>
    1.09
    [@BOS@]
    1.09
    <unused41>
    1.09
    <unused14>
    1.08
    <unused8>
    1.08
    <unused28>
    1.08
    <unused52>
    1.08
    <unused79>
    1.08
    <unused3>
    1.07
    Activations Density 0.227%

    No Known Activations