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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 103130
    Prev
    Next
    INDEX
    Explanations

    elements related to violent or risky scenarios involving humiliation and injury

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

    ancient masterpieces

    np_acts-logits-general · gemini-2.5-flash-lite

    mentions of explicit physical violence or bodily harm (words describing being attacked, injured, or subjected to violent actions).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0

    descriptions of harmful or abusive behavior—especially graphic violence, injury, dangerous chemical use, or sexual exploitation (notably involving children), often directed at a specific person.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
    ;->
    -0.36
    prefixer
    -0.34
     savez
    -0.34
    yakarta
    -0.34
     BaseModel
    -0.33
    migrationBuilder
    -0.32
    krä
    -0.32
    AxisAlignment
    -0.32
     Cientí
    -0.32
     ligiloj
    -0.31
    POSITIVE LOGITS
    MessageTagHelper
    0.60
    ſelf
    0.53
    featureID
    0.53
     myſelf
    0.51
    ंदीखरीदारी
    0.50
     мәкал
    0.49
    évaluateur
    0.49
    SourceChecksum
    0.49
     Meksiku
    0.48
    hitheatre
    0.48
    Activations Density 0.201%

    No Known Activations