© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11484
    Prev
    Next
    INDEX
    Explanations

    contrastive conjunctions across languages

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

    whereas and contrastive conjunctions

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

    whereas

    np_max-act-logits · claude-4-5-sonnet
    Triggered by @skarne
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    inely
    0.38
    🔩
    0.37
     ግን
    0.37
    ifères
    0.37
    ă
    0.37
    āla
    0.36
    也很
    0.36
    azim
    0.36
    也不是
    0.35
    νή
    0.35
    POSITIVE LOGITS
     whereas
    2.09
     sedangkan
    1.98
    whereas
    1.91
     tandis
    1.84
     Whereas
    1.64
     Sedangkan
    1.64
    Whereas
    1.58
     بينما
    1.56
    而
    1.49
     mentre
    1.48
    Activations Density 0.447%

    No Known Activations