© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130528
    Prev
    Next
    INDEX
    Explanations

    The neuron specializes in spotting modal/hypothetical auxiliaries (words like “would,” “could,” “might,” “expect,” etc.) that introduce conditional or speculative statements.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    would expect / think

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     kanyang
    -1.05
     spesso
    -1.02
     ibang
    -0.99
    报记者
    -0.99
    ierz
    -0.99
     inginkan
    -0.98
    rà
    -0.98
     často
    -0.96
    relse
    -0.95
     lainnya
    -0.94
    POSITIVE LOGITS
     would
    1.50
    Would
    1.30
     logical
    1.18
    would
    1.18
     serait
    1.12
     given
    1.08
     be
    1.05
    不应该
    1.05
    Surely
    1.05
    一个
    1.03
    Activations Density 0.033%

    No Known Activations