© 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-3-27B-IT
    5. 40-GEMMASCOPE-2-RES-262K
    7. 22030
    Prev
    Next
    INDEX
    Explanations

    He deciding or acting

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_40_width_262k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Known
    0.39
    Normally
    0.39
    事先
    0.39
     known
    0.38
     বলার
    0.38
     പറയ
    0.38
     نقول
    0.38
    ド
    0.38
     descripción
    0.37
    сона
    0.37
    POSITIVE LOGITS
     vowed
    0.57
    就可以
    0.55
    จะต้อง
    0.55
     должна
    0.52
     უნდა
    0.52
     сможет
    0.50
     WOULD
    0.49
     shouldn
    0.49
     reasoned
    0.48
     könne
    0.48
    Activations Density 0.004%

    No Known Activations