© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16150
    Prev
    Next
    INDEX
    Explanations

    diverse languages and scripts

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

    The neuron fires on semantically rich, content-bearing words—nouns, verbs, adjectives, and proper names—while ignoring common function words.

    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
     they
    -1.36
    ⎓
    -1.21
     montón
    -1.14
     llevo
    -1.13
     dieselbe
    -1.13
     hlavně
    -1.13
     stagn
    -1.12
     thier
    -1.12
     químicas
    -1.09
     workflows
    -1.05
    POSITIVE LOGITS
    pography
    1.40
    Karakter
    1.23
    쐈
    1.21
    produksi
    1.19
    NotSupported
    1.16
     nueces
    1.16
     ditem
    1.13
     Kleid
    1.11
     Eier
    1.10
     not
    1.09
    Activations Density 0.018%

    No Known Activations