© 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. 957
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on isolated capitalized tokens that serve as section or heading labels (e.g. single-word titles at the start of a block).

    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
     Riemannian
    0.52
    😽
    0.51
    拵
    0.51
     fermionic
    0.51
    😸
    0.50
     carelessness
    0.48
    同じ
    0.47
    kelig
    0.47
    。"
    0.47
     obtuse
    0.47
    POSITIVE LOGITS
    -
    0.79
    4
    0.70
    1
    0.70
    3
    0.68
    6
    0.68
    7
    0.67
    2
    0.66
    5
    0.62
    9
    0.62
    8
    0.62
    Activations Density 0.000%

    No Known Activations