© 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
    5. 40-GEMMASCOPE-2-RES-16K
    7. 9302
    Prev
    Next
    INDEX
    Explanations

    useful or helpful information

    np_acts-logits-general · gemini-2.5-flash-lite

    helpful and harmless

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-pt/resid_post/layer_40_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    responsibility
    0.51
    uating
    0.46
     vụ
    0.45
    अप
    0.45
     प्रकृति
    0.45
     옳
    0.44
     आरोपियों
    0.44
     प्राथमिकता
    0.43
    ುತ್ತಿದ್ದ
    0.43
    ủng
    0.42
    POSITIVE LOGITS
     useful
    1.24
     Useful
    1.20
    Useful
    1.20
     полез
    1.17
     utile
    1.16
     útil
    1.14
    useful
    1.11
     utiles
    1.09
     مفید
    1.09
     berguna
    1.03
    Activations Density 0.209%

    No Known Activations