© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 38473
    Prev
    Next
    INDEX
    Explanations

    original condition

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

    The neuron spotlights technical or descriptive words formed with “-ed” or “-ified” endings (i.e. past-participle/adjectival forms).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     чемпи
    -0.88
     náš
    -0.85
     padrão
    -0.84
     செய
    -0.83
    pheres
    -0.83
     dessa
    -0.81
     señaló
    -0.81
    豔
    -0.80
     reacción
    -0.79
    キル
    -0.77
    POSITIVE LOGITS
     intact
    1.98
     preserved
    1.78
     unaltered
    1.72
    そのまま
    1.66
     untouched
    1.63
     preserving
    1.51
     preserve
    1.48
     retain
    1.48
     preserves
    1.46
     undisturbed
    1.45
    Activations Density 0.075%

    No Known Activations