© 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. 10-GEMMASCOPE-RES-131K
    7. 90305
    Prev
    Next
    INDEX
    Explanations

    since, then, how, can

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

    The neuron fires on common second‐person and modal/request words (e.g. you, please, can’t, would, anything) used in questions or polite instructions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     good
    -2.16
     with
    -2.13
     this
    -1.93
     these
    -1.92
     because
    -1.89
     used
    -1.88
     time
    -1.85
     carefully
    -1.84
     for
    -1.80
     When
    -1.80
    POSITIVE LOGITS
     rano
    1.66
     différents
    1.57
     rése
    1.48
     modernes
    1.48
     praktisch
    1.46
    putting
    1.45
    一时间
    1.44
     teknis
    1.41
    extremely
    1.41
    ieme
    1.40
    Activations Density 0.134%

    No Known Activations