© 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. 22-GEMMASCOPE-RES-131K
    4. 73919
    Prev
    Next
    INDEX
    Explanations

    cited unpublished

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

    This neuron detects technical or domain‐specific noun phrases linked by prepositions (e.g. “absent from practice,” “contain spaces,” “citation of unpublished dispositions,” “images from Google”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     pourront
    -1.02
    suitable
    -0.95
    paying
    -0.92
     jacobs
    -0.90
    losing
    -0.89
    Trav
    -0.88
     comprimento
    -0.87
    льника
    -0.87
     úplně
    -0.86
     вернулся
    -0.86
    POSITIVE LOGITS
     any
    1.41
     cualquier
    1.34
     ANY
    1.27
     or
    1.18
    Before
    1.14
     Before
    1.13
     unusual
    1.11
     любых
    1.09
    enschutz
    1.09
    涉及
    1.07
    Activations Density 0.096%

    No Known Activations