© 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. 10-GEMMASCOPE-RES-131K
    4. 75608
    Prev
    Next
    INDEX
    Explanations

    role or roles

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

    The neuron fires on mentions of someone’s “role,” i.e. tokens referring to roles or positions.

    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
     in
    -2.23
     so
    -1.52
     discernible
    -1.52
     plethora
    -1.51
     endearing
    -1.47
     poignant
    -1.45
     but
    -1.43
     on
    -1.38
     interestingly
    -1.34
     got
    -1.32
    POSITIVE LOGITS
    an
    1.56
    me
    1.41
    thinkable
    1.34
    我們
    1.34
    on
    1.32
    厹
    1.32
     ſur
    1.31
     détru
    1.30
     焼
    1.30
     inade
    1.30
    Activations Density 0.042%

    No Known Activations