© 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. 1757
    Prev
    Next
    INDEX
    Explanations

    sinking or sunken

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

    The neuron fires on occurrences of “sink” and its variants (sinking, sunk, sunken, sank) in text.

    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
    瑧
    -2.34
    from
    -1.96
     寫
    -1.94
    窳
    -1.91
    Muchos
    -1.91
    彼が
    -1.91
    Descubre
    -1.91
    CURLOPT
    -1.91
    ING
    -1.91
    されない
    -1.91
    POSITIVE LOGITS
    ↵
    3.16
    ’
    2.25
     avec
    2.25
     με
    2.17
    ୌ
    1.88
    s
    1.84
     into
    1.75
    месте
    1.73
     Gegenteil
    1.73
    archiwizowane
    1.73
    Activations Density 0.009%

    No Known Activations