© 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. 66763
    Prev
    Next
    INDEX
    Explanations

    recursive data structures and operations

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

    The neuron is triggered by code‐specific tokens—e.g. punctuation, operators, underscores, parentheses, pipes, and programming keywords—rather than ordinary prose.

    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
     geöffnet
    -0.90
     communautaire
    -0.87
     אֶ
    -0.85
    шивания
    -0.83
     dormitorio
    -0.83
     herbe
    -0.79
    ]}>
    -0.79
    Πε
    -0.79
     SCHOL
    -0.78
    hör
    -0.78
    POSITIVE LOGITS
     again
    1.23
     further
    1.17
     other
    0.98
     remaining
    0.97
     from
    0.93
     Again
    0.89
     subsequent
    0.87
     less
    0.86
     חלק
    0.86
     continued
    0.85
    Activations Density 0.003%

    No Known Activations