© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 74543
    Prev
    Next
    INDEX
    Explanations

    truly unique or interesting descriptions

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

    The neuron fires on hyperbolic, promotional intensifier phrases—especially words like “truly,” “once,” “a/an” and similar superlative modifiers used to hype or emphasize (e.g. “truly a,” “once-in-a-lifetime,” “unique”).

    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
     mirando
    -1.85
    皱眉
    -1.74
     intentando
    -1.73
     sosteniendo
    -1.52
     tratando
    -1.51
     viendo
    -1.49
    usitis
    -1.48
    previously
    -1.48
     leyendo
    -1.47
     pidiendo
    -1.47
    POSITIVE LOGITS
     truly
    2.22
     will
    1.74
    是一
    1.55
     sincerely
    1.43
     sincero
    1.37
     which
    1.34
     would
    1.33
     crescer
    1.33
    真正
    1.32
     plastico
    1.32
    Activations Density 0.021%

    No Known Activations