© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34042
    Prev
    Next
    INDEX
    Explanations

    aspirational concepts followed by specification

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

    The neuron fires on words that appear as part of section or article titles and headlines—especially proper names, branded terms, and other standout “headline” tokens.

    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
     boister
    -1.63
    簠
    -1.62
     technisch
    -1.62
     anc
    -1.57
     horrid
    -1.56
     magnific
    -1.55
     advogado
    -1.52
     alho
    -1.49
     唯美
    -1.48
    蛱
    -1.48
    POSITIVE LOGITS
     If
    2.13
     get
    1.94
     if
    1.94
     give
    1.89
    If
    1.88
     provide
    1.84
     where
    1.77
     Because
    1.73
     does
    1.71
     this
    1.70
    Activations Density 0.051%

    No Known Activations