© 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. 31-GEMMASCOPE-2-RES-16K
    7. 3035
    Prev
    Next
    INDEX
    Explanations

    The neuron is firing most strongly on tokens at the ends of sentences—i.e. it marks sentence‐boundary/period positions.

    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
    只好
    0.83
     muß
    0.77
     великолеп
    0.77
     workmen
    0.77
     пять
    0.77
     numai
    0.76
    ;.
    0.75
     четыре
    0.74
     handicapped
    0.74
    􀂾
    0.74
    POSITIVE LOGITS
     reimag
    1.45
     impactful
    1.38
     nuanced
    1.33
     relatable
    1.28
     통해
    1.20
     overarching
    1.18
     neoliberal
    1.18
     problemat
    1.18
     interdiscipl
    1.16
     leveraging
    1.15
    Activations Density 0.369%

    No Known Activations