© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 127993
    Prev
    Next
    INDEX
    Explanations

    The neuron activates most strongly on the conjunction “but,” signaling contrastive “but” usages.

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

    but introduces contrast or reality

    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
    ParseException
    -1.02
     nehmen
    -1.02
     also
    -0.99
    :[],
    -0.99
     więc
    -0.92
     allerdings
    -0.90
     special
    -0.89
     también
    -0.88
    </h1>
    -0.86
     general
    -0.86
    POSITIVE LOGITS
    actually
    1.40
    Actually
    1.33
     réalité
    1.32
     actually
    1.31
     aslında
    1.30
    其實
    1.27
    真正的
    1.25
    其实
    1.16
    实际上
    1.09
     reality
    1.09
    Activations Density 0.052%

    No Known Activations