© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 73075
    Prev
    Next
    INDEX
    Explanations

    chemical compounds and foreign languages

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

    text in non-English languages, particularly Japanese.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron strongly activates on numeric tokens (including numbers with decimals and related measurement/unit strings).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     to
    -1.53
     but
    -1.47
    F
    -1.42
    P
    -1.41
    E
    -1.37
    D
    -1.35
     for
    -1.34
    T
    -1.30
    S
    -1.27
    M
    -1.25
    POSITIVE LOGITS
     célè
    1.41
     deko
    1.35
     jaunes
    1.33
     fidé
    1.30
    sprzed
    1.28
     nyel
    1.28
     adoro
    1.27
    Тел
    1.27
    çok
    1.26
     rése
    1.25
    Activations Density 0.011%

    No Known Activations