© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 75718
    Prev
    Next
    INDEX
    Explanations

    do not or if

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

    The neuron strongly activates on structural tokens—punctuation marks (periods, commas, parentheses, dashes) and numeric step indicators—i.e. the formatting and list‐format markers in the recipe.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     procha
    -1.19
     navires
    -1.16
    drawSprites
    -1.16
     būtų
    -1.15
     acheteurs
    -1.13
     planètes
    -1.13
     tirs
    -1.12
     tille
    -1.12
     atuais
    -1.10
    glad
    -1.09
    POSITIVE LOGITS
     you
    3.41
     this
    2.58
     if
    2.44
     then
    2.22
     don
    2.11
     or
    2.02
     make
    2.02
     You
    1.90
     If
    1.85
     do
    1.73
    Activations Density 0.022%

    No Known Activations