© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 33920
    Prev
    Next
    INDEX
    Explanations

    desire and preference

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

    This neuron detects words expressing personal sentiment or desire—especially first-person evaluations (e.g. “I want,” “regret,” “feel,” “saddened”).

    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
     even
    -1.09
     different
    -1.09
     after
    -1.04
     may
    -1.03
     can
    -0.98
     might
    -0.96
     will
    -0.96
     that
    -0.87
     so
    -0.85
     insist
    -0.84
    POSITIVE LOGITS
    want
    1.27
    取决
    1.20
    本当は
    1.16
    誒
    1.05
     手绘
    1.02
    ಇ
    1.02
    有沒有
    1.01
     optimale
    0.99
     botines
    0.99
    㗎
    0.98
    Activations Density 0.028%

    No Known Activations