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. 10907
    Prev
    Next
    INDEX
    Explanations

    names of political leaders and references to government-related events

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

    Preceding a minister or chancellor's name

    np_acts-logits-general · gemini-2.0-flash

    prime minister names

    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
     locals
    -0.55
     spotlight
    -0.55
    ViewFeatures
    -0.54
    万美元
    -0.53
    tubers
    -0.53
    onsored
    -0.53
    trice
    -0.51
    abestanden
    -0.51
    xchange
    -0.50
    vore
    -0.50
    POSITIVE LOGITS
     Cabinet
    1.10
     cabinet
    1.05
    Cabinet
    1.04
    cabinet
    1.01
     Prime
    0.91
    Prime
    0.81
     Cabinets
    0.78
     ministers
    0.77
     government
    0.77
     Government
    0.76
    Activations Density 0.193%

    No Known Activations