© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 10389
    Prev
    Next
    INDEX
    Explanations

    BOT token placeholder

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

    This neuron fires on instructional or metadata phrases—especially infinitival/imperative verbs (e.g. “To determine…,” “Put up by…,” “According to…”) that introduce how-to or attribution sentences.

    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
                        
    0.79
                       
    0.76
                            
    0.73
                      
    0.68
    oting
    0.67
          
    0.67
                             
    0.66
    <start_of_image>
    0.64
    ↵
    0.63
                   
    0.63
    POSITIVE LOGITS
     irgende
    0.90
     einzige
    0.89
     Karite
    0.87
     conspiring
    0.86
    tBleStatus
    0.86
     pebb
    0.85
    ผู้ชมในสนาม
    0.85
    Ⴉ
    0.85
     ulike
    0.84
     autoWatch
    0.84
    Activations Density 0.001%

    No Known Activations