© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129890
    Prev
    Next
    INDEX
    Explanations

    This neuron detects long runs of repeated punctuation (e.g. sequences of dashes or other separator characters used as section/divider lines).

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

    often irreverent or work it

    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
     każdy
    -1.19
     bemer
    -1.12
    SEMICOLON
    -1.10
    昕
    -1.09
     ofreciendo
    -1.09
     intelig
    -1.07
     chegar
    -1.06
    Soon
    -1.06
    bewerken
    -1.06
     ícone
    -1.05
    POSITIVE LOGITS
    1.15
     mathe
    1.14
     anta
    0.98
    。

    0.98
     ZAP
    0.98
     ß
    0.98
    udière
    0.97
     azure
    0.91
     inferno
    0.90
     billi
    0.89
    Activations Density 0.033%
    No Known Activations