© 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-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 33661
    Prev
    Next
    INDEX
    Explanations

    training hard

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

    The neuron detects bits of direct speech that issue instructions or commands—especially coaching‐style encouragement like “run faster,” “you can hear him,” and so on.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     btw
    -0.93
    此外
    -0.90
     skall
    -0.86
     Кроме
    -0.84
     folks
    -0.84
     bowiem
    -0.84
     segera
    -0.83
     color
    -0.82
     langt
    -0.82
     живе
    -0.81
    POSITIVE LOGITS
    好像
    1.10
     provavelmente
    1.08
    definitely
    1.06
     будто
    1.02
    お願いします
    1.02
     morals
    0.99
    有點
    0.97
    worldwide
    0.95
     ialah
    0.94
     ≈
    0.93
    Activations Density 0.012%

    No Known Activations