© 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. 67209
    Prev
    Next
    INDEX
    Explanations

    both

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

    The neuron is highly activated by common English function words (e.g. articles, prepositions, conjunctions) such as “the,” “and,” “to,” “at,” “from,” 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
     this
    -0.92
     obok
    -0.84
    をしました
    -0.79
    楣
    -0.79
     lecz
    -0.78
    ത്തി
    -0.77
     silicona
    -0.77
    計測
    -0.75
    kkor
    -0.75
    lund
    -0.74
    POSITIVE LOGITS
    Both
    1.49
    both
    1.38
     både
    1.33
     both
    1.23
     BOTH
    1.13
    どちらも
    1.09
    BOTH
    1.03
     sowohl
    1.03
     Both
    1.02
    ทั้ง
    1.01
    Activations Density 0.139%

    No Known Activations