© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130430
    Prev
    Next
    INDEX
    Explanations

    The neuron flags the key topical or informationally dense tokens (proper names, main nouns/verbs) that carry the document’s core content.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    additional extras

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    皚
    -0.99
     olabilir
    -0.98
    Boxscore
    -0.97
    ifier
    -0.95
    まれる
    -0.91
     VARIOUS
    -0.90
    Katalog
    -0.90
    ਫ਼
    -0.90
    çage
    -0.89
     ("%
    -0.88
    POSITIVE LOGITS
     ancillary
    1.54
     additional
    1.23
     extras
    1.16
     separate
    1.14
     auxiliary
    1.09
     dodatkowe
    1.00
    部分
    0.98
     accessories
    0.98
     kys
    0.98
     separately
    0.96
    Activations Density 0.091%

    No Known Activations