© 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. 1671
    Prev
    Next
    INDEX
    Explanations

    high-

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

    This neuron detects the leading words in hyphenated compound modifiers (e.g. “high-,” “long-,” “first-,” “all-” in compounds like high-speed, long-term, first-degree, all-in-one).

    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
    d
    1.17
    et
    1.09
    ed
    1.00
    на
    1.00
    j
    0.92
    elina
    0.91
    b
    0.89
    er
    0.87
    z
    0.85
    is
    0.80
    POSITIVE LOGITS
     practicable
    0.72
    eliers
    0.69
     transmittance
    0.68
     történ
    0.67
     philanthropic
    0.66
     logistical
    0.64
     kritis
    0.64
     stewardship
    0.64
     storey
    0.64
     invertebrate
    0.63
    Activations Density 0.021%

    No Known Activations