© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 122227
    Prev
    Next
    INDEX
    Explanations

    references to holders or devices that secure objects

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

    holders

    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
     Sewell
    -0.57
     Epif
    -0.55
    Personensuche
    -0.54
    addGap
    -0.54
     Meth
    -0.52
     Nw
    -0.52
     ​​
    -0.51
     Kansas
    -0.51
    CppMethod
    -0.50
     Pea
    -0.50
    POSITIVE LOGITS
    holder
    1.22
     Holder
    1.13
    Holder
    1.08
    HOLDER
    1.07
     holder
    1.02
    holders
    0.97
     Holders
    0.95
     holders
    0.86
    holding
    0.82
    hold
    0.78
    Activations Density 0.002%

    No Known Activations