© 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. 65189
    Prev
    Next
    INDEX
    Explanations

    betting offers

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

    The neuron fires strongly on the very first non‐trivial token of each new text segment—i.e. the lead word or number (often a proper noun or numeral) at the start of a paragraph or article.

    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
    -1.19
     like
    -1.13
    нейших
    -1.09
    atalos
    -1.05
     our
    -1.03
     most
    -1.02
     my
    -1.01
     circa
    -0.98
     I
    -0.94
    AttributeSet
    -0.94
    POSITIVE LOGITS
    Pendidikan
    1.11
    Ǎ
    1.09
    Postgres
    1.05
     phiếu
    1.04
     появилась
    1.01
     člověk
    1.00
    riotic
    0.98
    pendidikan
    0.98
    ⴷ
    0.97
     سيد
    0.97
    Activations Density 0.002%

    No Known Activations