© 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. 27601
    Prev
    Next
    INDEX
    Explanations

    ironically followed by description

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

    The neuron fires on discourse markers and adjectives that signal irony, paradox, or rhetorical contrast (e.g. “ironically,” “paradoxical,” “those who… also seem…”).

    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
     bahkan
    -1.18
    甚至是
    -1.02
    Herm
    -0.93
     bewertet
    -0.93
     entweder
    -0.91
     forklar
    -0.91
     nejen
    -0.91
    فقط
    -0.91
    Edd
    -0.90
     varying
    -0.89
    POSITIVE LOGITS
    தே
    0.93
     stessi
    0.93
     themselves
    0.90
     itself
    0.85
     ہی
    0.84
    وية
    0.80
     ઉ
    0.79
     prek
    0.79
     ironically
    0.77
    insley
    0.77
    Activations Density 0.033%

    No Known Activations