© 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-262K
    7. 22299
    Prev
    Next
    INDEX
    Explanations

    should

    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_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
    স
    0.54
     takže
    0.48
    达成
    0.47
    身材
    0.46
     מת
    0.46
    profiss
    0.46
    য়েছে
    0.45
    anske
    0.45
     osu
    0.45
     Zat
    0.45
    POSITIVE LOGITS
     should
    0.92
     ideally
    0.80
     shouldn
    0.79
     Should
    0.75
     Ideally
    0.73
    Ideally
    0.70
     powin
    0.70
     SHOULD
    0.69
     powinien
    0.68
     devraient
    0.67
    Activations Density 0.093%

    No Known Activations