© 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. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 74751
    Prev
    Next
    INDEX
    Explanations

    Denmark

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on the Italian toponyms “Danimarca” (Denmark) and “Copenaghen” (Copenhagen), i.e. it detects mentions of Denmark (and its capital) in Italian text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ylko
    -0.07
    idding
    -0.07
     Pittsburgh
    -0.07
    ología
    -0.07
    andez
    -0.07
    727
    -0.07
    (infile
    -0.06
     ImGui
    -0.06
    NodeId
    -0.06
    zcze
    -0.06
    POSITIVE LOGITS
     Denmark
    0.07
     Ary
    0.07
    غراف
    0.07
     Danish
    0.06
    unsafe
    0.06
     Emil
    0.06
     pornost
    0.06
     aalborg
    0.06
     Relative
    0.06
    	can
    0.06
    Activations Density 0.071%

    No Known Activations