© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 9728
    Prev
    Next
    INDEX
    Explanations

    beginning of phrases or names

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

    The neuron fires on metadata keys and section‐heading tokens (e.g. “keywords,” “mock,” “stub,” “test,” “History and profile,” “born”) in structured or semi-structured text.

    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
     over
    -1.14
    研发
    -0.96
     THE
    -0.94
    )||
    -0.94
    userRepository
    -0.93
     antigua
    -0.91
    -0.91
     Jú
    -0.90
    力で
    -0.90
    <0xF4>
    -0.89
    POSITIVE LOGITS
    
    1.10
     geleg
    1.02
     bezeich
    1.02
    每個
    1.01
     végétal
    0.98
    度
    0.95
    Modifica
    0.94
    ूल
    0.94
    Programma
    0.93
     può
    0.93
    Activations Density 0.000%

    No Known Activations