© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 14653
    Prev
    Next
    INDEX
    Explanations

    document sections, formatting, common words

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

    The neuron specializes in spotting numeric tokens—any digits or number‐like strings (e.g. version numbers, counts, percentages) in the 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
     wezen
    -1.33
    ⁉
    -1.10
     gezegd
    -1.09
     partecipare
    -1.07
     accedere
    -1.05
     entrare
    -1.02
    almaz
    -1.01
    😰
    -1.00
    昨日は
    -1.00
     signora
    -0.98
    POSITIVE LOGITS
     (
    1.30
     aka
    1.19
    aka
    1.12
    ఽ
    1.08
     www
    0.96
     extranjera
    0.96
    without
    0.96
     AKA
    0.95
    phazard
    0.94
    vs
    0.93
    Activations Density 0.092%

    No Known Activations