© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 68879
    Prev
    Next
    INDEX
    Explanations

    code-related identifiers

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

    The neuron fires on standalone name‐tokens inside code or markup—i.e. pieces of identifiers or labels (e.g. “next,” “Windows,” “right,” “Album,” “INFO,” etc.) rather than ordinary words.

    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
     még
    -1.06
     drei
    -1.02
    に戻る
    -1.00
     violación
    -0.99
     uñas
    -0.99
    oudoune
    -0.99
     perempuan
    -0.97
     revela
    -0.97
     mierda
    -0.97
    (!__
    -0.96
    POSITIVE LOGITS
     نیز
    1.24
    これも
    1.09
     troubling
    1.01
     εικό
    0.99
     first
    0.99
     which
    0.94
    が
    0.92
     scrivere
    0.92
     труда
    0.91
     subversive
    0.91
    Activations Density 0.039%

    No Known Activations