© 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. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 15231
    Prev
    Next
    INDEX
    Explanations

    Access rights, introduction, statistics, commands, diseases

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

    section headings followed by colons in document structures.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    This neuron fires strongly on colon tokens (":") typically used as separators in titles, headings, or section introductions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    toppers
    -1.53
    strator
    -1.52
    >>>>>
    -1.51
    knows
    -1.48
    rizona
    -1.46
    '%(
    -1.45
    🤑
    -1.43
     DOLLARS
    -1.43
    🤟
    -1.41
    vasive
    -1.40
    POSITIVE LOGITS
     our
    1.78
    </strong>
    1.53
     +
    1.52
     where
    1.42
    cipais
    1.33
     Störungen
    1.32
     /
    1.30
     Specifically
    1.30
     Erkrankungen
    1.29
     Krankheiten
    1.28
    Activations Density 0.060%

    No Known Activations