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. 20-GEMMASCOPE-RES-16K
    6. 31
    Prev
    Next
    INDEX
    Explanations

    representations of numerical values or mathematical structures

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

    long alphanumeric codes or identifiers, particularly those with mixed numbers and letters.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    numerical data, especially sequences of numbers and decimals.

    oai_token-act-pair · gpt-4oTriggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    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
    MessageOf
    -0.50
    transQ
    -0.50
     jsPsych
    -0.50
    InSection
    -0.48
    LookAnd
    -0.47
    tagHelperRunner
    -0.47
    følge
    -0.47
    ConstraintMaker
    -0.46
    ècie
    -0.45
    ⤹
    -0.45
    POSITIVE LOGITS
     <<<<<<<<<<<<<<
    0.38
     rootView
    0.36
    Jegyzetek
    0.35
     autorytatywna
    0.34
     CanadaChoose
    0.33
    Wie
    0.31
    Dans
    0.31
    enumii
    0.29
    iecie
    0.29
    Vice
    0.28
    Activations Density 0.245%

    No Known Activations