© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 12900
    Prev
    Next
    INDEX
    Explanations

    words like captain, track, raft

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

    The neuron fires most strongly on content‐bearing, relatively uncommon lexical tokens (i.e. salient nouns, adjectives, or proper names) rather than on high-frequency function words.

    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_24_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
     사람이
    0.74
    想
    0.74
     옳
    0.74
     또
    0.70
     Ủ
    0.67
    本
    0.66
    Histoire
    0.65
     아직
    0.65
    思
    0.64
     배우
    0.64
    POSITIVE LOGITS
     semidefinite
    0.84
     utama
    0.83
     geospatial
    0.82
     tertinggi
    0.78
     سادہ
    0.78
    ҳои
    0.77
     circadian
    0.77
     outlying
    0.76
     uniaxial
    0.75
     уні
    0.75
    Activations Density 0.000%

    No Known Activations