© 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. 65654
    Prev
    Next
    INDEX
    Explanations

    related to or part of

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

    The neuron consistently lights up on semantically important “content” words—nouns, adjectives, and verbs—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
    .)}
    -0.97
    ÑOS
    -0.96
     Bíblia
    -0.94
    ByteString
    -0.94
     =
    -0.94
    fir
    -0.92
     praticamente
    -0.92
     هل
    -0.91
     &&
    -0.91
    unos
    -0.90
    POSITIVE LOGITS
     or
    2.02
     some
    1.35
     may
    1.23
     might
    1.22
     hoặc
    1.16
     хотя
    1.16
     atleast
    1.15
     with
    1.14
     perhaps
    1.13
     because
    1.12
    Activations Density 0.136%

    No Known Activations