© 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. 33421
    Prev
    Next
    INDEX
    Explanations

    SQL script

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

    The neuron strongly activates on content words appearing in answer passages (i.e. after the “A:” marker) rather than on question text.

    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
    が
    -1.04
     adding
    -0.99
    gracias
    -0.90
    现在的
    -0.90
    íticos
    -0.87
    ・・。
    -0.84
    神話
    -0.83
     achieving
    -0.83
    最近は
    -0.82
    と
    -0.80
    POSITIVE LOGITS
    getYear
    1.20
    ⿻
    1.02
     sprzedaż
    1.01
    一抹
    0.98
     ilustraciones
    0.98
     międzynarod
    0.97
     persoonlijke
    0.96
    翳
    0.96
     belangrijke
    0.95
     sorpresas
    0.95
    Activations Density 0.041%

    No Known Activations