© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 16187
    Prev
    Next
    INDEX
    Explanations

    introduces explanations

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

    The neuron strongly activates on common verbs—including auxiliary and modal verbs (be/been, have, will, could) and their surrounding verbal forms (compared, classify, summarized, quantitatively)—marking general verb constructions rather than content words.

    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
    zzling
    -0.99
     verfolgen
    -0.94
    ådet
    -0.93
    hiers
    -0.93
    xically
    -0.93
    oirs
    -0.92
    Kegiatan
    -0.92
    sembles
    -0.90
    ishwa
    -0.90
     was
    -0.88
    POSITIVE LOGITS
     these
    1.70
    這些
    1.52
     they
    1.48
    这些
    1.47
     those
    1.47
     These
    1.41
     list
    1.40
     listed
    1.38
     them
    1.31
     эти
    1.25
    Activations Density 0.120%

    No Known Activations