© 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. 34282
    Prev
    Next
    INDEX
    Explanations

    debug model result Phone Q

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

    The neuron responds to tokens that form parts of file and directory names—especially sequences of digits, underscores, hyphens, and file‐extension fragments.

    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
     bedien
    -1.08
    masker
    -1.01
     estampado
    -1.00
     frasco
    -1.00
     oike
    -0.99
     reservado
    -0.98
    morph
    -0.98
    cropping
    -0.97
     or
    -0.96
    retirement
    -0.94
    POSITIVE LOGITS
     indor
    1.02
     प्रयास
    1.01
    ensdag
    0.96
     ranting
    0.96
    出来了
    0.96
     ぽ
    0.95
     peup
    0.95
    الله
    0.94
     cerveau
    0.91
    yně
    0.90
    Activations Density 0.008%

    No Known Activations