© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 13742
    Prev
    Next
    INDEX
    Explanations

    questions ending in punctuation

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

    The neuron responds to negation or contradiction words and dismissive interjections (e.g. “not,” “but,” “nonsense,” “although”).

    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
     greu
    -0.81
     באופן
    -0.80
     lavable
    -0.79
    Worse
    -0.79
    Unfortunately
    -0.78
     unfortunately
    -0.78
     uneasy
    -0.78
     Agence
    -0.77
     tikai
    -0.77
     ESPEC
    -0.77
    POSITIVE LOGITS
     despite
    1.10
     why
    1.00
     then
    0.96
    とは思
    0.95
     logic
    0.90
     argumento
    0.89
     argumentos
    0.87
     przecież
    0.84
     terwijl
    0.82
     argument
    0.82
    Activations Density 0.049%

    No Known Activations