© 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. 9370
    Prev
    Next
    INDEX
    Explanations

    air and associated concepts

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

    The neuron is tuned to detect the character sequence “air” (i.e. occurrences of “air” within words).

    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
     nerve
    0.67
    edges
    0.64
     MCG
    0.61
     çeşit
    0.59
    酤
    0.57
     triples
    0.56
     passam
    0.55
     ኢ
    0.55
    धर्म
    0.55
     Nerve
    0.55
    POSITIVE LOGITS
    Air
    1.02
     Air
    1.02
     air
    0.96
     vzduchu
    0.96
     एयर
    0.95
    air
    0.91
    空气
    0.86
    AIR
    0.83
     возду
    0.83
     空気
    0.81
    Activations Density 0.168%

    No Known Activations