© 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. 34-GEMMASCOPE-RES-131K
    7. 83976
    Prev
    Next
    INDEX
    Explanations

    paralleling, parallelized, paralytic

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

    The neuron activates on any occurrence of the subword “paral” (as in parallel, paralyzing, paralytic, parallax, 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_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
    zessin
    -0.74
     ucap
    -0.74
     retained
    -0.71
    異なります
    -0.69
     Gettysburg
    -0.69
    Kraj
    -0.68
    Ắ
    -0.68
    هما
    -0.68
     molecular
    -0.68
     normas
    -0.68
    POSITIVE LOGITS
     paralle
    0.98
     Paralle
    0.94
    paralle
    0.86
     paral
    0.82
    Parallel
    0.80
    lep
    0.77
     Parallel
    0.76
    ufige
    0.76
     parallèle
    0.75
     parallel
    0.73
    Activations Density 0.036%

    No Known Activations