© 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. 10-GEMMASCOPE-RES-131K
    4. 104005
    Prev
    Next
    INDEX
    Explanations

    hockey, fatty acids, command, Columbus, unit

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

    The neuron fires on single tokens that are part of formal technical, scientific, or institutional terms (e.g. “auto_conf,” “polyunsaturated,” “Biosphere,” “Strategic Command,” “GnRH,” “legislature”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     åren
    -1.61
     kado
    -1.55
     jawa
    -1.45
     tanga
    -1.42
     cyclisme
    -1.41
     dortmund
    -1.39
     yoox
    -1.39
     hawa
    -1.39
     iko
    -1.35
     errone
    -1.34
    POSITIVE LOGITS
     with
    2.31
     we
    1.93
     when
    1.84
     our
    1.74
    </h4>
    1.68
    ׇ
    1.64
     it
    1.61
     factores
    1.59
     tantangan
    1.58
     mesmo
    1.55
    Activations Density 0.143%

    No Known Activations