© 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. 40-GEMMASCOPE-2-RES-262K
    7. 26581
    Prev
    Next
    INDEX
    Explanations

    control the center; win duels

    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
     backdoor
    0.40
    谨慎
    0.40
     flaw
    0.38
     Conditional
    0.38
    廈
    0.37
    hmad
    0.37
     flaws
    0.37
    蟶
    0.37
    韌
    0.36
     Crystall
    0.36
    POSITIVE LOGITS
     battles
    0.60
     contests
    0.55
     contested
    0.55
     contest
    0.54
     fís
    0.52
     Battles
    0.50
     ruck
    0.50
     grapple
    0.50
     concursos
    0.50
     disputed
    0.49
    Activations Density 0.003%

    No Known Activations