© 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. 129319
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on words that denote the reversal or restoration of a prior state—terms like “lifted,” “cleared,” “free,” “back,” “reopened,” “restored,” or “reconstructed.”

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

    and continuation

    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
     bloquer
    -0.91
     destacado
    -0.90
     préc
    -0.88
     négy
    -0.86
     distinción
    -0.85
     oczeki
    -0.84
     européennes
    -0.83
     dépass
    -0.82
     maintenir
    -0.82
    alimentation
    -0.81
    POSITIVE LOGITS
     then
    1.38
     restored
    1.18
    重新
    1.15
     resumed
    1.08
     восстанов
    1.04
     still
    1.02
     resume
    0.97
     vẫn
    0.95
    再開
    0.94
     recovery
    0.92
    Activations Density 0.037%

    No Known Activations