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. 113845
    Prev
    Next
    INDEX
    Explanations

    directional indicators in mathematical contexts

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

    directional arrows and transformations

    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
    chymal
    -0.59
    ndice
    -0.58
     calendriers
    -0.54
    Климаты
    -0.54
    jabi
    -0.52
     }}$}
    -0.50
    isDebugEnabled
    -0.50
     Implants
    -0.49
     שוליים
    -0.49
    digarh
    -0.48
    POSITIVE LOGITS
    rightarrow
    1.39
    →
    0.95
     →
    0.85
    leftarrow
    0.79
    longrightarrow
    0.78
    mapsto
    0.75
     ->
    0.70
    return
    0.68
    Rightarrow
    0.66
    overrightarrow
    0.65
    Activations Density 0.018%

    No Known Activations