© 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. 22-GEMMASCOPE-RES-131K
    7. 65460
    Prev
    Next
    INDEX
    Explanations

    opposites, furthest, equivalents, range

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

    The neuron fires strongly on multi‐word noun phrases or compounds that convey measures, degrees, extremes, or oppositions (e.g. “opposite of,” “full spectrum,” “go-ahead goal,” “four horsemen”).

    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
     and
    -1.57
    KAZ
    -1.17
     nke
    -1.13
    èbres
    -1.03
     transferencia
    -1.03
     ------
    -1.02
     TALL
    -1.02
     واتس
    -1.02
     ……………………
    -1.02
     този
    -1.01
    POSITIVE LOGITS
     a
    1.82
     an
    1.22
     any
    1.10
    pre
    1.10
     More
    1.07
    𝐡
    1.05
     equivalent
    1.03
    mo
    1.03
    bu
    1.02
     Having
    1.00
    Activations Density 0.088%

    No Known Activations