© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 74968
    Prev
    Next
    INDEX
    Explanations

    properties of things

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

    The neuron strongly activates on very high‐frequency tokens—common function words (e.g. “to,” “be,” “can,” “have,” “a,” “with,” etc.) and other frequent short 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
     its
    -2.00
     it
    -1.25
     itself
    -1.21
     has
    -1.21
     is
    -1.10
     two
    -1.08
    它
    -0.99
    是个
    -0.93
    它的
    -0.92
    ꒷
    -0.90
    POSITIVE LOGITS
     themselves
    2.02
     they
    1.45
    themselves
    1.42
     których
    1.25
     которые
    1.20
     are
    1.13
     eivät
    1.09
    their
    1.06
     They
    1.05
    They
    1.03
    Activations Density 0.083%

    No Known Activations