© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 3570
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on imperative “action” words in code comments or docs (e.g. Create, Specify, Set, Keep, Dump, Try), i.e. instruction-style verbs used to describe programming steps.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
    <strong>
    0.91
    <b>
    0.83
    Adv
    0.75
    𝐋
    0.68
    LAND
    0.68
    Land
    0.67
    Maurice
    0.66
    ANY
    0.66
    ○○
    0.66
    Экс
    0.66
    POSITIVE LOGITS
     stuff
    1.15
     variables
    1.11
     nécessaire
    1.05
     nécessaires
    1.05
     కోసం
    1.04
     Stuff
    1.04
     pentru
    1.02
     Variables
    1.00
     Vars
    1.00
     necesitamos
    1.00
    Activations Density 0.007%

    No Known Activations