© 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. 7331
    Prev
    Next
    INDEX
    Explanations

    equal rights. by attracting. bearded seals. Unilever consistently. licenses to transmit

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

    This neuron fires strongest on the last word of each text segment (i.e. it marks segment endings).

    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
    ﻨ
    0.85
    ﺘ
    0.83
    ﺨ
    0.76
    reversed
    0.72
    膳
    0.72
    旗下
    0.71
    ﺤ
    0.69
    ﺖ
    0.69
    足以
    0.68
    ିକ
    0.68
    POSITIVE LOGITS
     hunger
    0.76
     heat
    0.75
     visible
    0.75
     brilliant
    0.74
     dedication
    0.74
     heats
    0.74
     foundry
    0.73
     visualization
    0.72
     dues
    0.72
    اء
    0.72
    Activations Density 0.000%

    No Known Activations