© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25100
    Prev
    Next
    INDEX
    Explanations

    distribute holy water, build sand, attach connectors

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

    The neuron strongly activates on the word “to,” especially when it introduces infinitive verbs or prepositional phrases.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     BIBSYS
    -1.16
     RSSSF
    -1.14
    殮
    -1.12
     embreagem
    -1.12
    伀
    -1.07
     полноцен
    -1.06
     '&#
    -1.05
    μμ
    -1.04
     [''],
    -1.04
     transferência
    -0.98
    POSITIVE LOGITS
    Often
    1.48
     often
    1.45
     their
    1.41
    Sometimes
    1.34
     Often
    1.31
     such
    1.27
    Usually
    1.22
    Their
    1.20
     sometimes
    1.14
     Sometimes
    1.13
    Activations Density 0.070%

    No Known Activations