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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 122896
    Prev
    Next
    INDEX
    Explanations

    instances of the word "conduct" in various forms

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    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
    Spoljašnje
    -0.59
     épar
    -0.58
     Socialista
    -0.57
     Италијани
    -0.54
    totalPrice
    -0.53
     ErrorMessage
    -0.51
     Água
    -0.49
    smar
    -0.48
     sourire
    -0.47
     newValue
    -0.47
    POSITIVE LOGITS
     Conduct
    1.00
     conducted
    0.94
     conduct
    0.93
    conduct
    0.91
     Conducted
    0.88
    Conduct
    0.85
     CONDUCT
    0.82
     conducting
    0.79
    conducted
    0.79
     conducts
    0.78
    Activations Density 0.017%

    No Known Activations