© 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. 15890
    Prev
    Next
    INDEX
    Explanations

    contrastive conjunctions or punctuation

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

    The neuron flags punctuation and sentence‐structuring tokens—especially commas, numbered labels, citation markers, and common discourse connectors (e.g. “then,” “because,” “but,” “and”).

    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
     sessão
    -1.02
    voie
    -1.02
     összefoglal
    -1.01
     Generale
    -1.00
    hause
    -1.00
    чня
    -0.99
     cérami
    -0.98
     boden
    -0.96
    quele
    -0.96
     quidem
    -0.96
    POSITIVE LOGITS
     which
    1.02
     or
    0.94
    孰
    0.90
    昙
    0.85
    ReusableCell
    0.84
     lovely
    0.83
    ᅠ
    0.82
    cephala
    0.82
    ify
    0.81
     diversa
    0.81
    Activations Density 0.109%

    No Known Activations