© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 36171
    Prev
    Next
    INDEX
    Explanations

    scoring points

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

    The neuron detects numeric score details and point‐margin expressions (e.g. numbers paired with “point,” “advantage,” “lead,” “run,” etc.) in sports game descriptions.

    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
    bumptech
    -1.12
    tachten
    -1.09
     RISERV
    -1.05
    temun
    -1.05
    moga
    -1.01
    versario
    -1.01
    terapia
    -1.00
    kalender
    -1.00
    komend
    -0.98
    Already
    -0.97
    POSITIVE LOGITS
    ῖς
    1.05
     led
    1.04
    ans
    1.02
     prowad
    1.01
    問い
    0.99
    常用的
    0.94
     control
    0.93
    (
    0.92
    pare
    0.91
    調理
    0.91
    Activations Density 0.009%

    No Known Activations