© 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. 116698
    Prev
    Next
    INDEX
    Explanations

    foreign words and technical terms

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

    The neuron flags words in a user’s direct request for help or instructions—first-person pronouns (“I,” “me”), request verbs (“help,” “instruct,” “do,” “show,” “run”), and the “to” of “how to/what to” constructions.

    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
     also
    -1.07
     would
    -0.91
     képes
    -0.88
    ffilm
    -0.87
     Лей
    -0.86
     when
    -0.86
     mxArray
    -0.85
     zwięks
    -0.85
     membersihkan
    -0.85
     ALSO
    -0.84
    POSITIVE LOGITS
     compromis
    1.13
     toro
    1.04
    ländischen
    1.03
     autonomie
    1.03
     diamant
    1.00
     aquare
    0.99
     frigor
    0.98
    
    0.98
     infal
    0.98
     bril
    0.97
    Activations Density 0.001%

    No Known Activations