© 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. 22200
    Prev
    Next
    INDEX
    Explanations

    evaluative nouns

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

    The neuron fires on single, salient nouns naming objects or titles (e.g. “movie,” “show,” “item,” “site,” “service”), flagging mentions of key entities.

    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
     a
    -2.27
     famoso
    -2.09
     célèbre
    -1.73
    même
    -1.69
    ଃ
    -1.47
    ⢅
    -1.46
     his
    -1.45
     allerlei
    -1.45
    tersebut
    -1.44
    
    -1.44
    POSITIVE LOGITS
    They
    1.73
     evidently
    1.59
     for
    1.54
     took
    1.52
     chose
    1.50
     intense
    1.49
     encourage
    1.43
     intensely
    1.41
     lengthy
    1.41
    Also
    1.39
    Activations Density 0.021%

    No Known Activations