© 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. 3123
    Prev
    Next
    INDEX
    Explanations

    replace with

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

    The neuron fires on occurrences of the word “replace” (and its inflected forms like “replacing” or “replaced”), i.e. verbs indicating substitution of one thing for another.

    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
    ال
    1.05
    (',',$
    0.98
    planes
    0.84
    ित
    0.82
    ري
    0.81
    ल
    0.79
    ூக
    0.79
     Adj
    0.79
     巻
    0.79
    绎
    0.79
    POSITIVE LOGITS
     reempl
    1.21
     Replacing
    1.15
     заменить
    1.15
     substituting
    1.11
     remplacement
    1.11
     replace
    1.09
     substituted
    1.09
     sostit
    1.09
     substituir
    1.08
     vervangen
    1.08
    Activations Density 0.042%

    No Known Activations