© 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. 31-GEMMASCOPE-2-RES-16K
    4. 1288
    Prev
    Next
    INDEX
    Explanations

    copy and paste

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

    The neuron strongly activates on the conjunction “and.”

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

    the word "and" when it connects two parallel actions, clauses, or items in a sequence.

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

    the coordinating conjunction used to link words, phrases, or clauses, especially when chaining actions or paired items.

    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_31_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
    しかし
    0.56
    А
    0.53
    M
    0.52
     그러나
    0.52
    с
    0.50
     一
    0.48
                   
    0.48
     પરંતુ
    0.47
    ただし
    0.47
        
    0.46
    POSITIVE LOGITS
     whatnot
    0.83
     sebagainya
    0.79
    rogens
    0.68
    ouille
    0.67
     zároveň
    0.64
    romeda
    0.64
    rog
    0.61
    ंगाबाद
    0.60
     wondered
    0.59
    rews
    0.58
    Activations Density 0.889%

    No Known Activations