© 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. 4115
    Prev
    Next
    INDEX
    Explanations

    can/could indicate possibility

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

    The neuron is picking up auxiliary/modal verbs (e.g. can, could, would, may, had) that signal a reviewer’s personal stance or recommendation.

    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
     Telekom
    0.66
    このように
    0.64
    🦳
    0.62
     Archived
    0.60
     Broadcast
    0.59
     Istituto
    0.59
    Broadcast
    0.58
     TELE
    0.58
     그러면
    0.58
     staked
    0.58
    POSITIVE LOGITS
    了不少
    0.79
     düşünüyorum
    0.67
    otry
    0.66
     lotta
    0.63
    了很多
    0.63
    ową
    0.62
    生素
    0.62
    lectricité
    0.62
     lluvias
    0.62
    जिसे
    0.60
    Activations Density 0.159%

    No Known Activations