© 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. 10-GEMMASCOPE-RES-131K
    4. 70665
    Prev
    Next
    INDEX
    Explanations

    would followed by verb

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

    The neuron activates strongly on the modal verb “would.”

    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
    is
    -1.76
     of
    -1.57
    </h1>
    -1.51
     increases
    -1.49
    </h2>
    -1.48
     or
    -1.42
     multifaceted
    -1.39
     increased
    -1.38
    りますが
    -1.36
    g
    -1.36
    POSITIVE LOGITS
    ⤒
    1.79
    iffance
    1.67
     Its
    1.53
     różnych
    1.48
    
    1.46
    蜮
    1.44
    athons
    1.44
    にほんブログ村
    1.38
     
    1.38
    biendo
    1.36
    Activations Density 0.060%

    No Known Activations