© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 14394
    Prev
    Next
    INDEX
    Explanations

    restore, redeem, regain, reset

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

    The neuron activates on phrases that state responsibilities, obligations, or commitments to resolve issues (e.g. delivery times, disclaimers of liability, “will communicate” and “make every effort to resolve”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    什么
    1.17
    на
    1.13
    가능
    0.96
    𝒟
    0.93
    有點
    0.91
    个
    0.89
    每一個
    0.88
    每一个
    0.87
    ோ
    0.86
    Cuál
    0.85
    POSITIVE LOGITS
     corrective
    1.50
     corrections
    1.48
     восста
    1.45
     reinstatement
    1.42
     refund
    1.41
     restore
    1.41
    重新
    1.40
    回来
    1.39
     віднов
    1.37
     correction
    1.35
    Activations Density 5.480%

    No Known Activations