© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 130188
    Prev
    Next
    INDEX
    Explanations

    high accuracy, excellent quality

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

    The neuron flags descriptive degree words—adjectives and adverbs that indicate high or notable quality, extent, or frequency (e.g. “easy,” “high,” “excellent,” “often,” “frequently”).

    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
     ilumina
    -1.80
    驺
    -1.77
    ↵
    -1.55
     drau
    -1.53
     cheia
    -1.52
     zweimal
    -1.50
     It
    -1.48
     mediano
    -1.48
     there
    -1.48
     exclusivamente
    -1.48
    POSITIVE LOGITS
    8
    2.13
    6
    2.03
    3
    1.97
    豳
    1.84
    7
    1.84
    9
    1.84
    ότη
    1.79
     elbise
    1.73
    as
    1.70
    ástica
    1.70
    Activations Density 0.005%

    No Known Activations