© 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. 22-GEMMASCOPE-RES-131K
    4. 67914
    Prev
    Next
    INDEX
    Explanations

    clean room, clean up, clean slate

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

    The neuron is primarily triggered by the substring “clean” (in words like clean, cleanable, clean_up, nested_clean, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
          
    -2.33
    锨
    -2.33
     ряда
    -2.27
     clínico
    -2.25
     "
    -2.20
     необходимо
    -2.19
     üblichen
    -2.17
     Placing
    -2.16
     gewünschten
    -2.16
     reciben
    -2.14
    POSITIVE LOGITS
    鯪
    2.84
    ༛
    2.77
    »:
    2.77
    »;
    2.59
    蟶
    2.55
    媜
    2.48
    ...»
    2.39
     стоковые
    2.38
    Ꮄ
    2.36
    оргаш
    2.31
    Activations Density 0.013%

    No Known Activations