© 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. 73593
    Prev
    Next
    INDEX
    Explanations

    type_context, type_species, CMD_REQUEST

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

    The neuron fires on tokens containing underscores—particularly parts of snake_case or underscore‐delimited identifiers (e.g. “type_…”).

    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
    s
    -2.33
    人是
    -2.22
     So
    -1.92
    ↵↵
    -1.90
    ского
    -1.87
     sarebbe
    -1.84
    柃
    -1.83
    IANGLES
    -1.80
     пол
    -1.80
     $*
    -1.80
    POSITIVE LOGITS
    嚳
    2.55
     adoro
    2.44
    癰
    2.30
    ⢪
    2.20
    躄
    2.19
    しなければ
    2.09
    慥
    2.09
     bezpiecze
    2.09
     имеют
    2.09
     solchen
    2.05
    Activations Density 0.054%

    No Known Activations