© 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. 64777
    Prev
    Next
    INDEX
    Explanations

    but suffice to say

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

    The neuron strongly activates on sentence-breaking punctuation (commas, parentheses) and high-weight discourse pivots like “but,” marking clause or sentence transitions.

    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
    ==========
    -0.90
    ANDRA
    -0.86
    χι
    -0.86
    zesi
    -0.85
    WireFormatLite
    -0.85
     Hendrik
    -0.83
     không
    -0.83
     BERLIN
    -0.82
     Manifesto
    -0.82
    спасибо
    -0.80
    POSITIVE LOGITS
     suffice
    1.28
    Suff
    1.23
     what
    1.05
     Suff
    1.05
     лишь
    1.00
    Instead
    0.98
     hanya
    0.90
    かというと
    0.90
    suff
    0.90
     Instead
    0.88
    Activations Density 0.024%

    No Known Activations