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. 22452
    Prev
    Next
    INDEX
    Explanations

    references to contracts or agreements

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

    Function words, punctuation marks, and technical terms that serve as connecting elements or specific terminology in legal, business, and technical documentation.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    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
    NEW
    -0.38
    __*/
    -0.38
    NewLabel
    -0.37
    roek
    -0.35
     Glob
    -0.35
    vább
    -0.35
    PageRoute
    -0.35
    tilizer
    -0.35
    ources
    -0.35
    通り
    -0.35
    POSITIVE LOGITS
     periodically
    0.73
     intermittently
    0.56
     continually
    0.50
     irgendwann
    0.50
    whenever
    0.49
     whenever
    0.49
     constantly
    0.48
    IsMutable
    0.48
    principalTable
    0.47
     constantemente
    0.47
    Activations Density 0.156%

    No Known Activations