© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 70341
    Prev
    Next
    INDEX
    Explanations

    explaining how something is

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

    The neuron fires on the word “how” when used as a discourse marker (e.g. “that’s how …,” “that’s how we saw it,” etc.) to introduce or explain something.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    儼
    -1.57
     milano
    -1.57
     récemment
    -1.52
     hann
    -1.44
     palio
    -1.44
     "
    -1.41
    
    -1.41
    ровке
    -1.38
     verona
    -1.38
    韪
    -1.38
    POSITIVE LOGITS
     when
    1.70
    䜣
    1.47
    ki
    1.44
    matchCondition
    1.42
     way
    1.41
    el
    1.40
    ku
    1.40
    ruary
    1.40
     pentru
    1.39
    flexGrow
    1.34
    Activations Density 0.014%

    No Known Activations