© 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. 15718
    Prev
    Next
    INDEX
    Explanations

    difficulties or things to avoid

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

    The neuron is essentially flagging content-bearing, relatively rare or salient words (uncommon nouns, adjectives, or verbs) rather than common 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
     convergent
    0.65
    দীর্ঘ
    0.63
     victories
    0.61
     pertandingan
    0.60
    ウォッチ
    0.60
     joyous
    0.58
    とともに
    0.57
     converg
    0.57
    每一个
    0.56
     heartfelt
    0.56
    POSITIVE LOGITS
    不敢
    1.44
    害怕
    1.30
     afraid
    1.19
     слишком
    1.17
    不愿意
    1.16
    不允许
    1.16
    不想
    1.14
    怕
    1.14
     terlalu
    1.13
    太多
    1.12
    Activations Density 1.061%

    No Known Activations