© 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. 34383
    Prev
    Next
    INDEX
    Explanations

    setting parameters to values

    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
     தான்
    0.41
    ഘ
    0.40
     गुजर
    0.39
     Nah
    0.38
     thorough
    0.37
     ಸಂಪ
    0.37
     Hoe
    0.36
     Ek
    0.36
    付出
    0.36
    move
    0.36
    POSITIVE LOGITS
     accordingly
    0.63
    为
    0.61
    Accordingly
    0.51
    為
    0.51
    的值
    0.50
     priorities
    0.49
     ablaze
    0.49
     corretamente
    0.48
     appropriately
    0.47
     parameters
    0.47
    Activations Density 0.016%

    No Known Activations