© 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. 31-GEMMASCOPE-2-RES-16K
    7. 529
    Prev
    Next
    INDEX
    Explanations

    introduces contrast

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

    The neuron strongly activates on coordinating conjunctions or discourse-linking words (e.g. "but," "and," "also"), especially when they start a clause.

    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_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
     ,
    0.40
                
    0.39
                    
    0.39
    0.39
    =>
    0.39
    。
    0.38
     =>
    0.38
    ↵↵
    0.37
                                
    0.36
    AP
    0.36
    POSITIVE LOGITS
     has
    0.64
     transcends
    0.62
     deserves
    0.61
     exudes
    0.61
     provides
    0.59
     requires
    0.59
     encompasses
    0.59
     caters
    0.59
     precludes
    0.59
     could
    0.58
    Activations Density 0.693%

    No Known Activations