© 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. 84384
    Prev
    Next
    INDEX
    Explanations

    roof rails

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

    The neuron detects common multi‐word collocations (frequently co‐occurring function‐word sequences or idiomatic phrase chunks).

    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
     отличный
    -1.14
    TEMA
    -1.11
    廡
    -1.09
     wikipedia
    -1.05
     ayrıca
    -1.04
     Dinas
    -1.03
     Guayaquil
    -1.02
     ue
    -1.02
     consistente
    -1.02
     VERT
    -1.01
    POSITIVE LOGITS
     in
    1.37
     TWO
    1.23
     with
    1.21
    ！(
    1.16
     which
    1.16
     as
    1.13
     from
    1.12
     could
    1.11
     through
    1.10
     lavish
    1.07
    Activations Density 0.102%

    No Known Activations