© 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. 10-GEMMASCOPE-RES-131K
    7. 39448
    Prev
    Next
    INDEX
    Explanations

    work areas and directories

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

    This neuron fires on technical compound phrases centered on “work” or “working” (e.g. “work documents,” “working memory,” “WORK DIR,” “work Tree”).

    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
    ’
    -3.44
    ruguay
    -2.19
    '
    -2.17
     buque
    -2.03
     helicóptero
    -1.84
    './
    -1.81
     muut
    -1.80
     \
    -1.79
     menyen
    -1.77
     desapar
    -1.76
    POSITIVE LOGITS
    啻
    2.31
    2.03
     Any
    1.98
    齜
    1.98
     These
    1.97
     zweiten
    1.95
     Work
    1.95
     augmenté
    1.92
     estos
    1.88
     intrepid
    1.83
    Activations Density 0.025%

    No Known Activations