© 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. 7061
    Prev
    Next
    INDEX
    Explanations

    without followed by a consequence

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

    This neuron specifically detects occurrences of the word “without” (in any case) as a key negative/disjunctive cue.

    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
     طول
    0.73
     مللي
    0.70
    だけ
    0.69
    দুই
    0.67
    dq
    0.66
    뭡
    0.66
    时长
    0.64
     শুধ
    0.63
    что
    0.63
     дли
    0.62
    POSITIVE LOGITS
     regard
    1.06
     any
    0.97
     an
    0.93
     adanya
    0.91
     resorting
    0.91
     being
    0.87
     prejudice
    0.84
     recourse
    0.82
     a
    0.80
     intending
    0.80
    Activations Density 0.112%

    No Known Activations