© 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-3-270M
    5. 12-GEMMASCOPE-2-RES-16K
    7. 7726
    Prev
    Next
    INDEX
    Explanations

    subtracting from both sides

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-270m-pt/resid_post/layer_12_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
    Francis
    1.14
     CCTV
    1.11
     harbours
    1.07
     infallible
    1.04
     Francis
    1.01
     manned
    0.99
    Charles
    0.97
    铯
    0.94
     peduncul
    0.92
     durée
    0.90
    POSITIVE LOGITS
     Simplifying
    2.14
     canceling
    2.13
    Combine
    2.08
     straightening
    2.07
     Simplify
    2.06
     subtracting
    2.03
    simplify
    2.02
     Elim
    2.01
    Simplify
    2.01
     Elimination
    1.99
    Activations Density 0.003%

    No Known Activations