© 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. 34-GEMMASCOPE-RES-131K
    4. 103358
    Prev
    Next
    INDEX
    Explanations

    as if / as though

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

    The neuron specifically detects occurrences of the word “as” when it’s used in comparative or hypothetical constructions (e.g. “as if,” “as though”).

    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
     kanske
    -1.21
     enligt
    -1.13
     alltså
    -1.08
     istället
    -1.07
     incendios
    -1.07
     conformidad
    -1.06
    Когда
    -1.05
    kologi
    -1.04
    来越
    -1.03
     samtidigt
    -1.03
    POSITIVE LOGITS
     if
    6.94
     though
    4.75
    if
    3.58
     если
    3.56
     اگر
    2.91
    though
    2.89
     If
    2.78
    Though
    2.77
    If
    2.64
     якщо
    2.59
    Activations Density 0.030%

    No Known Activations