© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72767
    Prev
    Next
    INDEX
    Explanations

    you can see or hear

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

    The neuron fires on second-person references—particularly “you” and its common contractions or future‐tense forms (e.g. you’ll, you’ve, won’t).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    𝗴
    -1.18
    erdings
    -1.15
     comemor
    -1.13
    क़
    -1.13
     lú
    -1.09
     Archivado
    -1.08
     legenda
    -1.05
     kys
    -1.04
     wp
    -1.02
     officier
    -1.02
    POSITIVE LOGITS
     see
    2.06
     heard
    1.71
     seeing
    1.67
     thấy
    1.66
     saw
    1.65
     seen
    1.61
    เห็น
    1.55
     দেখ
    1.38
     sees
    1.37
    Seeing
    1.23
    Activations Density 0.054%

    No Known Activations