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
    4. Residual Stream - 16k
    5. 21-GEMMASCOPE-RES-16K
    6. 6048
    Prev
    Next
    INDEX
    Explanations

    terms associated with programming constructs and data structures

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.hook_resid_post
    Hook Layer
    21
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    <unused52>
    -0.98
    <unused16>
    -0.98
    <unused28>
    -0.98
    <unused14>
    -0.98
    <unused68>
    -0.97
    <unused8>
    -0.97
    <unused79>
    -0.97
    <unused41>
    -0.96
    <unused23>
    -0.96
    [@BOS@]
    -0.96
    POSITIVE LOGITS
    ?
    0.32
     perna
    0.31
    <eos>
    0.31
    .
    0.29
     necesidad
    0.28
     explique
    0.27
     posesión
    0.27
    ↵↵
    0.27
     ninguno
    0.27
     regalías
    0.26
    Activations Density 0.143%

    No Known Activations