Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 5086
    Prev
    Next
    INDEX
    Explanations

    financial implications and governmental budgetary concerns

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

    Government revenue or funds

    np_acts-logits-general · gemini-2.0-flash

    public treasury funds

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    principalTable
    -0.85
    存于互联网档案馆
    -0.79
     ModelExpression
    -0.78
    AndEndTag
    -0.76
    Hochspringen
    -0.69
    FieldNumber
    -0.69
    IntoConstraints
    -0.68
    ThroughAttribute
    -0.68
    jgl
    -0.68
    ➟
    -0.68
    POSITIVE LOGITS
     taxpayers
    1.02
     coffers
    0.85
     treasury
    0.82
     taxpayer
    0.81
     Treasury
    0.73
     pockets
    0.67
     devlet
    0.67
     state
    0.67
     valsty
    0.65
     State
    0.65
    Activations Density 0.267%

    No Known Activations