© 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. 34-GEMMASCOPE-RES-131K
    7. 75518
    Prev
    Next
    INDEX
    Explanations

    Clayton, Microsoft, Clostridium, Policy

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

    The neuron activates on capitalized tokens—especially proper names, place names, acronyms, and other uppercase identifiers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Accepted
    -0.82
     permis
    -0.81
    paraíso
    -0.79
    licaciones
    -0.79
     выделя
    -0.79
    鄰
    -0.77
    accepted
    -0.77
     conventions
    -0.75
     excess
    -0.75
    crom
    -0.74
    POSITIVE LOGITS
    Clay
    0.90
     Clay
    0.89
     Clayton
    0.88
    CLAY
    0.84
     clay
    0.79
     clays
    0.79
    Clayton
    0.76
    }}$,
    0.73
    /+
    0.73
    voy
    0.73
    Activations Density 0.015%

    No Known Activations