© 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. 130752
    Prev
    Next
    INDEX
    Explanations

    This neuron detects sentence boundaries, firing strongly on the end-of-sentence period token and the following sentence-initial token.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    said

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     as
    -1.59
     at
    -1.52
     perplexing
    -1.50
     intensely
    -1.48
     supremely
    -1.41
     undeniably
    -1.40
     lavish
    -1.38
     exquisitely
    -1.38
     unrelenting
    -1.34
     Making
    -1.33
    POSITIVE LOGITS
    probably
    1.28
    镲
    1.26
    不太
    1.23
    perhaps
    1.21
    不是很
    1.20
     ........
    1.17
    ꔹ
    1.17
     verhind
    1.16
    ówki
    1.15
     mantenga
    1.12
    Activations Density 0.018%

    No Known Activations