© 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. 22-GEMMASCOPE-RES-131K
    7. 35608
    Prev
    Next
    INDEX
    Explanations

    formal document terms

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

    The neuron is sensitive to formal procedural or administrative terminology—words like “informed,” “consent,” “warranty,” “apologizes,” “transcribed,” etc., that occur in official or methodological statements.

    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
     by
    -1.15
     allerede
    -1.15
     that
    -1.11
     through
    -1.10
    Proč
    -1.08
     tæn
    -1.08
    mayacak
    -1.08
    Também
    -1.05
     intptr
    -1.05
     an
    -1.04
    POSITIVE LOGITS
    ఝ
    1.15
     multicolore
    1.13
    則是
    1.10
     bobina
    1.07
     새
    1.05
    则是
    1.05
     bleus
    1.04
     entrer
    1.03
     epoca
    1.02
    采用了
    1.02
    Activations Density 0.080%

    No Known Activations