© 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. 130725
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric and formal named‐entity tokens (years, measurements, statistics, and capitalized proper nouns) in biographical or factual contexts.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    Entities and their common follow-ups

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     that
    -1.49
    kaan
    -1.09
    內容簡介
    -1.07
     SwitchCompat
    -1.07
    €”
    -1.06
     which
    -1.05
    lň
    -1.05
    ഞ്
    -1.04
    neiden
    -1.04
    yaret
    -1.02
    POSITIVE LOGITS
    unjukan
    1.22
     several
    1.14
    dsquared
    1.14
    iorgio
    1.14
     العديد
    1.13
     einwand
    1.12
     عندما
    1.11
     incluyen
    1.04
     seguenti
    1.03
    だけでなく
    1.00
    Activations Density 0.133%

    No Known Activations