© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 306
    Prev
    Next
    INDEX
    Explanations

    The neuron activates primarily on hyphen (“-”) tokens connecting 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
    -
    0.43
     multi
    0.42
     ٢
    0.41
     de
    0.41
     $*$-
    0.41
     beurre
    0.41
    -”
    0.41
     –
    0.40
     ١
    0.40
     lvl
    0.40
    POSITIVE LOGITS
    based
    0.60
    এর
    0.57
    related
    0.56
    sized
    0.56
    прежнему
    0.53
    oriented
    0.52
    themed
    0.51
    centered
    0.50
    type
    0.49
    shaped
    0.48
    Activations Density 1.170%

    No Known Activations