© 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. 130307
    Prev
    Next
    INDEX
    Explanations

    The neuron flags meta‐speech cues—words like “remark,” “references,” “joked,” “glib,” etc.—that signal the author is making an aside or humorous comment.

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

    said, joking

    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
    currentValue
    -0.84
     вър
    -0.84
     crunch
    -0.82
     and
    -0.81
    viz
    -0.81
     Lettres
    -0.81
     Gestion
    -0.79
     tiny
    -0.77
    ory
    -0.77
    maj
    -0.75
    POSITIVE LOGITS
     może
    1.01
     poderia
    1.00
    ならば
    1.00
     تعدى
    0.97
    damen
    0.97
    Gdy
    0.97
    kapital
    0.96
    JScrollPane
    0.96
    Może
    0.94
    原始内容存档于
    0.93
    Activations Density 0.019%

    No Known Activations