© 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-3-27B-IT
    5. 41-GEMMASCOPE-2-TRANSCODER-262K
    7. 65997
    Prev
    Next
    INDEX
    Explanations

    not good, incorrect, or messy

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_41_width_262k_l0_small_affine
    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
     मुं
    0.38
    듐
    0.38
     விய
    0.38
    ("/")
    0.36
     Tiles
    0.36
    Tiles
    0.35
    泮
    0.35
    toRadians
    0.34
    FileLine
    0.34
    砚
    0.33
    POSITIVE LOGITS
     sloppy
    1.70
     poorly
    1.67
     improperly
    1.35
     careless
    1.34
     incompetent
    1.32
     carelessness
    1.31
     slo
    1.27
     shod
    1.26
     inept
    1.25
     carelessly
    1.25
    Activations Density 0.213%

    No Known Activations