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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 29685
    Prev
    Next
    INDEX
    Explanations

    references to male subjects

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

    he [verb]

    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
     in
    -0.33
    όμε
    -0.32
    vspace
    -0.32
     грн
    -0.30
    ỡ
    -0.29
     Natural
    -0.28
    ぷ
    -0.28
     of
    -0.28
     across
    -0.27
    tic
    -0.27
    POSITIVE LOGITS
    He
    1.00
     Ellos
    1.00
     ſhe
    0.99
    they
    0.99
    THEY
    0.98
    Mereka
    0.96
    she
    0.95
    They
    0.94
     he
    0.94
     ſeine
    0.93
    Activations Density 0.086%

    No Known Activations