© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 9207
    Prev
    Next
    INDEX
    Explanations

    movement directions (up, down, left, right)

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

    The neuron fires most strongly on isolated tokens that occur in formal metadata lines—things like affiliation statements, official codes, addresses, exam-center labels, or other structured administrative labels.

    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
    ↯
    1.00
    ๚
    0.95
    
    0.92
    Созда
    0.88
     haga
    0.88
    keyList
    0.85
    发送
    0.84
    我知道
    0.84
    ל
    0.84
    ସ
    0.84
    POSITIVE LOGITS
     recirc
    0.68
    o
    0.68
    oes
    0.67
     J
    0.67
     d
    0.66
    ે
    0.66
    oit
    0.64
    oise
    0.64
     ZO
    0.63
    erver
    0.63
    Activations Density 0.000%

    No Known Activations