© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 7363
    Prev
    Next
    INDEX
    Explanations

    just wanted to send a reminder

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

    The neuron is primarily triggered by the contrastive conjunction “but.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    the word "but" in contexts where the author is expressing a contrasting or qualifying statement.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    conversational hedges and discourse markers in first‑person comments, especially polite asides, apologies, thanks, and contrastive pivots.

    oai_token-act-pair · gpt-5Triggered 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
     decides
    0.65
     teniendo
    0.64
     கூட
    0.63
    ’
    0.63
    !
    0.61
     memang
    0.61
     !,
    0.60
     usando
    0.59
     ràng
    0.59
    !”,
    0.59
    POSITIVE LOGITS
    Sug
    0.61
    MESSAGE
    0.61
    砂糖
    0.56
    Нау
    0.56
    枻
    0.56
    СР
    0.55
    を発
    0.55
    स्ट्रेल
    0.55
    čki
    0.54
    inė
    0.54
    Activations Density 0.081%

    No Known Activations