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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 1103
    Prev
    Next
    INDEX
    Explanations

    code syntax involving declarations and function setups

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

    Code-related terms and class names

    np_acts-logits-general · gemini-2.0-flash

    specialized code blocks

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    TagMode
    -0.40
     saurait
    -0.37
    gså
    -0.30
    XmlType
    -0.29
    FunctionFlags
    -0.28
    pkl
    -0.27
    cljs
    -0.26
    ckså
    -0.26
    FC
    -0.26
     canlynol
    -0.25
    POSITIVE LOGITS
    RegressionTest
    0.71
    Personendaten
    0.69
    ArgsConstructor
    0.65
    RectangleBorder
    0.63
    脚注の使い方
    0.61
    ReusableCell
    0.60
    nodoc
    0.59
    Personensuche
    0.59
    InstanceState
    0.56
    EndInit
    0.56
    Activations Density 0.094%

    No Known Activations