© 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. 22-GEMMASCOPE-RES-131K
    7. 72158
    Prev
    Next
    INDEX
    Explanations

    citations and legal references

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

    This neuron activates on numeric tokens—sequences of digits such as page numbers, years, volumes, and other numeric entries in reference citations.

    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
     from
    -1.05
    years
    -0.98
     when
    -0.90
    ñones
    -0.90
    Legături
    -0.85
    Each
    -0.84
    ffiche
    -0.83
    راب
    -0.82
     what
    -0.81
     whose
    -0.81
    POSITIVE LOGITS
     porque
    0.85
     tež
    0.82
     potrà
    0.82
     certeza
    0.80
    anueva
    0.78
     titulaire
    0.78
    動き
    0.78
     obligator
    0.78
    corpor
    0.77
     reuni
    0.77
    Activations Density 0.004%

    No Known Activations