© 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. 31-GEMMASCOPE-2-RES-16K
    7. 11965
    Prev
    Next
    INDEX
    Explanations

    erotic content refusal

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

    The neuron strongly activates on structural break tokens (blank‐line or paragraph separators), i.e. it detects paragraph or scene boundaries.

    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_31_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
     Zhang
    0.62
     Xiao
    0.57
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.56
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.55
     Xia
    0.55
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.54
     gratifying
    0.54
     말이
    0.54
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.53
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.53
    POSITIVE LOGITS
    Nonetheless
    0.75
     Nonetheless
    0.66
    Apart
    0.65
    とはいえ
    0.65
    could
    0.64
    Ɛ
    0.64
    checking
    0.61
    whom
    0.61
    Checking
    0.59
    ซิตี
    0.59
    Activations Density 0.012%

    No Known Activations