© 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. 16209
    Prev
    Next
    INDEX
    Explanations

    empty or not filled

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

    The neuron is tuned to detect numeric values and number‐related tokens (digits, measurements, and counts).

    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
    Connexion
    -0.83
     サラダ
    -0.81
    CFU
    -0.78
    vorg
    -0.77
    ☽
    -0.75
    mög
    -0.71
    ələ
    -0.71
    əsi
    -0.69
     moderation
    -0.69
     crowded
    -0.69
    POSITIVE LOGITS
     empty
    1.31
     lonely
    1.28
     idle
    1.21
    Empty
    1.20
     boredom
    1.19
    isEmpty
    1.18
     isEmpty
    1.17
    empty
    1.16
     loneliness
    1.14
     barren
    1.13
    Activations Density 0.052%

    No Known Activations