© 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. 10-GEMMASCOPE-RES-131K
    7. 39342
    Prev
    Next
    INDEX
    Explanations

    but, but later, but failed, but moved, but in, but star

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

    The neuron detects contrastive discourse markers—especially “but” and “however” (often with a following comma).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ner
    -1.82
    end
    -1.61
    ],
    -1.59
     (
    -1.52
    int
    -1.49
    l
    -1.48
    za
    -1.45
    )
    -1.41
    nt
    -1.41
    ar
    -1.40
    POSITIVE LOGITS
    adalah
    1.87
    gruss
    1.66
    hermoso
    1.65
     mooi
    1.62
     zü
    1.61
     prachtig
    1.56
    detek
    1.55
     vriende
    1.55
    jauh
    1.52
     mencari
    1.50
    Activations Density 0.136%

    No Known Activations