© 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. 22-GEMMASCOPE-RES-131K
    4. 27925
    Prev
    Next
    INDEX
    Explanations

    clauses introduced by "as"

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

    The neuron strongly fires on the comma (and the key verb immediately before it) that ends an introductory clause at the start of a sentence—i.e. it marks the boundary after an opening “As …,” or similar lead‐in phrase.

    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
     when
    -2.20
     since
    -1.48
     after
    -1.36
     if
    -1.27
     після
    -1.21
     когда
    -1.20
     before
    -1.18
     wanneer
    -1.16
     had
    -1.16
     where
    -1.15
    POSITIVE LOGITS
     also
    1.35
     increasingly
    1.33
     jednocześnie
    1.27
     более
    1.26
     more
    1.24
    越来越
    1.24
     simultaneously
    1.22
     também
    1.20
     gradually
    1.18
     Simultaneously
    1.16
    Activations Density 0.037%

    No Known Activations