© 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. 28493
    Prev
    Next
    INDEX
    Explanations

    sentence beginnings in various languages

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

    The neuron fires strongly on tokens at the beginnings and ends of sentences, i.e. it is detecting sentence boundaries.

    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
     then
    -1.94
     there
    -1.71
     also
    -1.70
     також
    -1.38
     также
    -1.37
     can
    -1.34
     também
    -1.30
     other
    -1.25
     many
    -1.21
     others
    -1.21
    POSITIVE LOGITS
    不但
    1.34
    だけでなく
    1.27
     rasch
    1.25
     nejen
    1.25
    大丈夫です
    1.15
     decidi
    1.12
    Ꭾ
    1.12
     ด้วย
    1.12
     grà
    1.10
    不仅
    1.09
    Activations Density 0.098%

    No Known Activations