© 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. 95612
    Prev
    Next
    INDEX
    Explanations

    episode, names, numbers, busy states

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

    The neuron responds to specific identifiers—especially numbers, dates, episode counts, and proper names—i.e. tokens that denote named entities or numeric references.

    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
     glow
    -1.02
     sqlSession
    -1.02
     such
    -0.95
     कुछ
    -0.93
    templateUrl
    -0.93
    */;
    -0.92
    datable
    -0.92
    ᵃ
    -0.92
     can
    -0.90
     most
    -0.90
    POSITIVE LOGITS
     intenção
    1.08
     zvyš
    1.03
     umożli
    1.02
     asumir
    0.99
    つも
    0.98
     tasarım
    0.98
     reveló
    0.96
     cantit
    0.95
     revelou
    0.95
    utschein
    0.95
    Activations Density 0.000%

    No Known Activations