© 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. 10-GEMMASCOPE-RES-131K
    4. 75503
    Prev
    Next
    INDEX
    Explanations

    preposition followed by noun or verb

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

    The neuron flags words that serve as procedural step‐ or section‐heading cues—especially gerund/participle forms (numbing, introducing, boarding, confirming) and other header‐style tokens (Warm-up, Embarkation, Registration) that mark stages in instructions, itineraries, or tutorials.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     such
    -2.11
     that
    -2.08
     those
    -1.96
     you
    -1.73
     this
    -1.71
     That
    -1.55
     not
    -1.54
     used
    -1.45
     these
    -1.45
     when
    -1.45
    POSITIVE LOGITS
     nimmer
    1.34
     soucis
    1.34
     ulub
    1.30
     ervaring
    1.28
     ложка
    1.27
    Ꮚ
    1.26
     voyant
    1.26
     Comité
    1.25
     lavable
    1.24
     rêver
    1.23
    Activations Density 0.084%

    No Known Activations