© 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. 66965
    Prev
    Next
    INDEX
    Explanations

    example:

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

    This neuron fires on the structural markers of formal mathematical definitions or theorems—especially punctuation (commas and periods) and keywords like “property,” as well as capitalized words that begin new clauses or sentences.

    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
     UnityEditor
    -0.99
     registre
    -0.91
     comerci
    -0.90
     inova
    -0.89
     bronz
    -0.89
    hermoso
    -0.86
     حالات
    -0.85
     enthousi
    -0.85
    日閲覧
    -0.84
    städ
    -0.84
    POSITIVE LOGITS
     this
    1.63
    This
    1.30
     such
    1.25
     такого
    1.17
     هذه
    1.14
     этого
    1.05
     diesem
    1.05
    こうした
    1.04
    様な
    1.04
    Such
    1.04
    Activations Density 0.033%

    No Known Activations