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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 37386
    Prev
    Next
    INDEX
    Explanations

    references to personal behavior and interactions

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

    phrasal verb completions

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

    phrases that describe deliberate, agent-driven actions or plans in narrative or instructional contexts (e.g., “I/he/you” doing, deciding, starting, cutting, highlighting, or directing).

    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
     Luego
    -0.31
     cielos
    -0.29
    rawtypes
    -0.28
     com
    -0.27
     constater
    -0.27
     North
    -0.27
     kontakt
    -0.26
     dessus
    -0.25
     sahip
    -0.24
     Nachfolger
    -0.24
    POSITIVE LOGITS
    IntoConstraints
    0.79
    tagHelperRunner
    0.77
    StructEnd
    0.77
    CppMethod
    0.73
     Exacts
    0.71
     ModelExpression
    0.70
    DockStyle
    0.70
    Chham
    0.69
    évaluateur
    0.67
     laſſen
    0.65
    Activations Density 0.048%

    No Known Activations