© 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. 27963
    Prev
    Next
    INDEX
    Explanations

    speeches and statements expressing opinions

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

    The neuron strongly activates on first‐person self‐references—e.g. “I,” “me,” “I think,” “I would like to…,” or “we” when the speaker is expressing personal intent or opinion.

    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
    ductors
    -1.41
    formatics
    -1.30
    signia
    -1.16
     menggel
    -1.14
    izzati
    -1.09
     '',
    -1.09
     ancaman
    -1.09
    loggen
    -1.06
    mapStateToProps
    -1.05
    //}
    -1.04
    POSITIVE LOGITS
     释放
    1.42
     briefly
    1.38
     again
    1.32
    大丈夫です
    1.26
    ύ
    1.20
     modestly
    1.15
     Anliegen
    1.14
    eseorang
    1.13
     acessar
    1.12
     australiano
    1.11
    Activations Density 0.006%

    No Known Activations