© 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. 22-GEMMASCOPE-RES-131K
    4. 569
    Prev
    Next
    INDEX
    Explanations

    mathematical or scientific context markers

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

    This neuron fires on comparison prompts—especially “Which is greater/smaller…?” style questions comparing two numeric expressions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     cucina
    -1.33
     sánchez
    -1.12
     svart
    -1.09
    шивка
    -1.08
     Luar
    -1.07
     guitare
    -1.06
     Réalisation
    -1.06
     people
    -1.06
     compagnon
    -1.06
     prevents
    -1.05
    POSITIVE LOGITS
     VERY
    1.30
     and
    1.23
     hierzu
    1.17
     tumultuous
    1.15
     fervent
    1.12
     marvelous
    1.12
     witty
    1.11
     매우
    1.10
     whimsical
    1.10
     strikingly
    1.10
    Activations Density 0.000%

    No Known Activations