© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 14494
    Prev
    Next
    INDEX
    Explanations

    The neuron detects discourse-level transition words—especially contrastive or continuative conjunctions (e.g. “however,” “but,” “so”)—that link or pivot between statements.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    笑容
    0.59
    vital
    0.56
     लची
    0.55
     ప్రభావ
    0.55
    लक्ष
    0.54
     लक्ष्य
    0.54
     அதிகரிக்கிறது
    0.53
    ELEMENT
    0.52
    высо
    0.52
    သောအ
    0.52
    POSITIVE LOGITS
     моего
    0.61
     তোমার
    0.61
     Photoshop
    0.60
     Пи
    0.60
     내가
    0.60
     OneDrive
    0.60
     நான்
    0.59
     লে
    0.59
     నేను
    0.58
     จะ
    0.58
    Activations Density 0.066%

    No Known Activations