© 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. 31-GEMMASCOPE-2-RES-65K
    7. 15086
    Prev
    Next
    INDEX
    Explanations

    exception handling

    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_31_width_65k_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
     irritability
    0.42
     détecter
    0.42
     pesky
    0.41
    原因是
    0.40
     irritated
    0.39
     //'
    0.39
    lattice
    0.39
     psychologically
    0.38
    ErrorInfo
    0.38
     irritating
    0.38
    POSITIVE LOGITS
    inals
    0.42
     adat
    0.41
     sử
    0.40
    onder
    0.40
     i
    0.38
     yalnız
    0.38
     ل
    0.38
    forming
    0.37
     dolu
    0.37
     making
    0.37
    Activations Density 0.062%

    No Known Activations