© 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. 79850
    Prev
    Next
    INDEX
    Explanations

    specific technical terms

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

    The neuron strongly responds to isolated identifier tokens—single‐letter or short names of variables, functions, types, or platforms (e.g. “object”, “X”, “f”, “u”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    specialized technical or academic jargon and notation, especially math symbols and programming-related tokens.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    closing parentheses in technical or scientific text.

    oai_token-act-pair · claude-4-5-sonnetTriggered 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
     on
    -1.42
     or
    -1.19
     because
    -1.11
     when
    -1.05
     by
    -1.05
     Unlike
    -1.01
     after
    -1.00
     and
    -0.99
    when
    -0.98
     in
    -0.96
    POSITIVE LOGITS
     somewhat
    1.08
     revered
    1.02
     THROUGH
    0.99
     lauded
    0.98
     viola
    0.96
     mehrerer
    0.96
     overly
    0.95
    dolce
    0.95
     vorgenommen
    0.94
    さみ
    0.94
    Activations Density 0.038%

    No Known Activations