© 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. 24-GEMMASCOPE-2-RES-16K
    7. 14882
    Prev
    Next
    INDEX
    Explanations

    End of sentence punctuation

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

    The neuron is signaling how far into the document you are, i.e. it activates most strongly on tokens appearing deep (late) in the text.

    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
     имеется
    0.87
     நன்கு
    0.84
     หาก
    0.82
     می‌باشد
    0.79
     ควร
    0.78
     Mayweather
    0.77
     รอ
    0.77
     ambos
    0.75
     takeout
    0.75
     считаю
    0.75
    POSITIVE LOGITS
     genom
    0.63
    कैसे
    0.61
    这就是
    0.60
     stolet
    0.59
    しかも
    0.59
     sogen
    0.58
    变成了
    0.58
    原来
    0.57
    成了
    0.57
     éton
    0.57
    Activations Density 0.175%

    No Known Activations