© 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. 39454
    Prev
    Next
    INDEX
    Explanations

    describing people and their roles

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

    The neuron is especially sensitive to gerund or present‐participle verbs (e.g. “live,” “growing,” “being”) that introduce personal experiences, roles, or life‐in‐context 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
    foundland
    -1.14
     ベー
    -1.05
    ↵↵↵↵↵↵↵↵↵↵↵↵↵
    -1.03
    rogenic
    -1.02
    是要
    -1.02
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    -1.01
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    -1.00
     zamiast
    -0.99
     vorgestellt
    -0.99
     metri
    -0.99
    POSITIVE LOGITS
    Pozri
    1.30
     in
    1.25
     ilyen
    1.23
     процес
    1.17
    Belum
    1.15
     anpass
    1.10
    richText
    1.09
    ladı
    1.09
     összefoglal
    1.09
     an
    1.08
    Activations Density 0.054%

    No Known Activations