© 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. 10597
    Prev
    Next
    INDEX
    Explanations

    instructions, making them more complex, diverse, and challenging. This involves adding constraints, requiring more steps, or demanding specific reasoning patterns. * **Iterative Expansion:** This process is repeated multiple times,the possibility of embedding clauses within clauses, and using complex sentence structures to connect multiple ideas, even if those ideas are only loosely related, and employing various rhetorical devices to maintain grammatical coherence, and remembering that the true test of a sentence isVlogs:** Tu vida diaria, viajes, experiencias. * **Educación:** Tutoriales, clases, explicaciones. * **Comedia:** Sketches, parodias, reacciones.complex subject instructions

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

    The neuron detects mentions of downloadable content or file‐format links (e.g. “Download,” “PDF,” “File,” “Excel”).

    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
              
    0.77
            
    0.73
             
    0.73
          
    0.73
    숑
    0.67
         
    0.66
                    
    0.65
     unclean
    0.64
    ayos
    0.63
    olulu
    0.61
    POSITIVE LOGITS
     grabado
    0.72
    ប្រភេទ
    0.70
     profund
    0.70
     dort
    0.69
     scru
    0.68
     simplement
    0.67
     thức
    0.66
     chiave
    0.66
     graça
    0.66
     ouv
    0.65
    Activations Density 0.000%

    No Known Activations