© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 36664
    Prev
    Next
    INDEX
    Explanations

    actions to achieve goals

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

    The neuron strongly activates on single-word verbs that denote providing, supplying, or facilitating something (e.g., cover, fuel, record, provide, improve).

    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
     trône
    -1.51
     aperçu
    -1.39
     aktivi
    -1.36
    raeli
    -1.33
    omsday
    -1.32
     Codable
    -1.30
    ebly
    -1.30
    modium
    -1.30
     telles
    -1.27
     trots
    -1.27
    POSITIVE LOGITS
     and
    2.19
     Pfarrer
    1.45
    produzione
    1.38
    umblr
    1.29
     There
    1.27
     pojawia
    1.27
     náš
    1.25
    pios
    1.23
     miała
    1.21
     an
    1.20
    Activations Density 0.043%

    No Known Activations