© 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. 34-GEMMASCOPE-RES-131K
    7. 108665
    Prev
    Next
    INDEX
    Explanations

    not be, year after, comes out

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

    The neuron responds strongest to common high‐frequency English function words—especially modal and auxiliary verbs (e.g. may, can, would, be, have) and other stop words.

    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
     either
    -1.20
     likely
    -1.06
     attempts
    -1.00
     indirectly
    -0.98
    可能
    -0.97
     whatever
    -0.96
     similar
    -0.96
     neither
    -0.96
    自己的
    -0.94
    俛
    -0.94
    POSITIVE LOGITS
     чёр
    0.98
     suivie
    0.93
    ⦑
    0.92
    ditch
    0.91
     Xx
    0.91
    𝕒
    0.90
    novu
    0.89
    conomie
    0.89
    xC
    0.87
    กรณ์
    0.86
    Activations Density 0.291%

    No Known Activations