© 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. 36026
    Prev
    Next
    INDEX
    Explanations

    either or

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

    The neuron responds most strongly to uncommon, low‐frequency tokens (i.e. unusually rare words and numbers) in the 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
     recens
    -1.32
     Egyp
    -1.31
     gelatina
    -1.30
     vocale
    -1.29
     maría
    -1.27
     dictionnaire
    -1.27
     infrar
    -1.25
     palco
    -1.23
     durer
    -1.23
     organisme
    -1.20
    POSITIVE LOGITS
     it
    1.40
     It
    1.30
     before
    1.18
     Still
    1.16
     Even
    1.15
     any
    1.15
     Both
    1.15
     I
    1.13
     because
    1.13
     When
    1.12
    Activations Density 0.056%

    No Known Activations