© 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. 22-GEMMASCOPE-RES-131K
    7. 68438
    Prev
    Next
    INDEX
    Explanations

    scientific study components

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

    scientific methodology terminology in research papers, particularly related to experimental design, apparatus, and data collection.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron selectively detects terms naming experimental methods or study‐design elements (e.g., apparatus, assay, recruitment, design).

    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
     mere
    -1.05
     判断
    -1.00
     doesn
    -0.95
     if
    -0.94
    多吃
    -0.93
     {
    -0.92
     nothing
    -0.91
    izzazioni
    -0.91
     simply
    -0.90
     Fragen
    -0.89
    POSITIVE LOGITS
     used
    1.36
     for
    1.21
     setup
    1.13
    甦
    1.08
     was
    1.08
    พาะ
    1.08
    곳
    1.05
    ֶּ
    1.01
    álise
    0.98
    mær
    0.98
    Activations Density 0.021%

    No Known Activations