© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 11861
    Prev
    Next
    INDEX
    Explanations

    environment characteristics

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

    The neuron fires on those little cue‐words that introduce or point to sources, attributions, and reference metadata—e.g. “labelled,” “with,” “from,” “Retrieved,” “Targets,” “cross,” “offer,” “obtaining,” etc.—marking citation-style or dataset-reference phrases.

    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
    SE
    0.93
    ry
    0.93
    sa
    0.86
    sp
    0.84
    kms
    0.84
    Kya
    0.84
    i
    0.84
    IL
    0.83
    ES
    0.83
    muy
    0.82
    POSITIVE LOGITS
     Disorder
    0.88
     イエロー
    0.79
     Atomic
    0.75
     AppCompatTheme
    0.74
     Metabolism
    0.74
     Revision
    0.73
     Disabilities
    0.72
     Parts
    0.70
     Testosterone
    0.70
     Device
    0.70
    Activations Density 0.000%

    No Known Activations