© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 7453
    Prev
    Next
    INDEX
    Explanations

    Which followed by a noun

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

    The neuron strongly activates on sentence-initial interrogative prompts—especially the word “Which”—marking the start of a question.

    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
    ות
    1.16
    綃
    1.16
    ларга
    1.14
     whatnot
    1.11
    ї
    1.06
    fellow
    1.04
    за
    0.98
    ো
    0.96
     elucidate
    0.94
    UnaryOperation
    0.93
    POSITIVE LOGITS
    되었다
    1.26
    吒
    1.26
     কিনা
    1.18
    くらいの
    1.17
    يد
    1.15
    koľ
    1.15
     sadde
    1.13
    ید
    1.12
    件事情
    1.12
     setores
    1.12
    Activations Density 0.030%

    No Known Activations