© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 1311
    Prev
    Next
    INDEX
    Explanations

    masked language modeling

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

    The neuron activates on past‐tense or past‐participle forms—i.e. words ending in “-ed.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-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
    Not in Any Lists

    No Comments

    Negative Logits
     الموافق
    0.60
     Allocation
    0.59
    iesz
    0.56
    0.55
     tác
    0.54
     Koles
    0.54
     Bộ
    0.54
     ा
    0.53
     Oleh
    0.53
     余
    0.52
    POSITIVE LOGITS
     up
    0.86
    在一起
    0.86
     versie
    0.79
    起来
    0.75
     şekilde
    0.74
    起來
    0.73
    ependent
    0.70
     appropriately
    0.70
     with
    0.68
     properly
    0.68
    Activations Density 0.685%

    No Known Activations