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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 116026
    Prev
    Next
    INDEX
    Explanations

    Boost Software License

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

    This neuron fires on code‐style punctuation runs (e.g. comment markers like “//”, directive hashes like “#”, long sequences of whitespace or slashes), i.e. tokens composed mainly of repeated punctuation.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    μως
    -0.83
    ghana
    -0.82
     Jahrhunder
    -0.82
    ageable
    -0.80
    )')
    -0.78
     standoff
    -0.75
    /.../
    -0.73
    RACE
    -0.73
    我国
    -0.71
     matchmaking
    -0.71
    POSITIVE LOGITS
     bezpiecze
    0.75
     dégustation
    0.73
    Vik
    0.73
    宮崎
    0.71
     Otra
    0.71
     exemption
    0.71
    来的
    0.70
    捗
    0.69
     sartén
    0.69
    銀座
    0.69
    Activations Density 0.003%

    No Known Activations