© 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. 100455
    Prev
    Next
    INDEX
    Explanations

    referring to

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

    The neuron activates on patent-style numeric references (e.g. figure numbers, reference numerals, and other numbered citations) in the text.

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

    references to figures and formal patent-specification phrasing that describe elements “with reference to” drawings and numerals.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    phrases that introduce cross-references to figures or other parts of technical patent documents.

    oai_token-act-pair · claude-4-5-sonnetTriggered 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
     он
    -1.83
    </
    -1.82
    {
    -1.80
    (!__
    -1.79
    によって
    -1.71
     was
    -1.59
    )
    -1.57
    でも
    -1.52
     of
    -1.52
     его
    -1.51
    POSITIVE LOGITS
    🫶
    1.49
     there
    1.45
    อื่น
    1.45
     ervaren
    1.43
    各大
    1.35
    Waar
    1.34
     relics
    1.32
     certificat
    1.30
     entrées
    1.29
    薷
    1.28
    Activations Density 0.053%

    No Known Activations