© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2906
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on numbered or otherwise enumerated list markers and section‐heading tokens (e.g. “Saturday:”, patent claim numbers, bullet‐style numerals).

    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
     socalled
    1.48
     wellknown
    1.26
     longterm
    0.81
    »،
    0.80
    ׳
    0.79
     ẞ
    0.73
    "،
    0.73
     ܀
    0.72
    ,»
    0.72
    িয়৷
    0.71
    POSITIVE LOGITS
      
    3.48
       
    2.23
        
    2.11
    <0xE3>
    1.91
          
    1.71
         
    1.68
            
    1.57
    1.49
     ​​
    1.38
              
    1.38
    Activations Density 1.014%

    No Known Activations