Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 84499
    Prev
    Next
    INDEX
    Explanations

    forum answers

    np_max-act · gemini-2.0-flash

    The neuron responds to polite request markers—especially “por favor” (and equivalents like “please”)—signaling that the user is asking politely (e.g. “por favor, envíame el código…”).

    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
     roční
    -0.06
    มาก
    -0.06
     Advice
    -0.06
    InputBorder
    -0.06
    uploaded
    -0.06
     zoek
    -0.06
     filtr
    -0.06
     Furniture
    -0.06
    .dropout
    -0.06
     Addition
    -0.06
    POSITIVE LOGITS
    .FLOAT
    0.07
    sembl
    0.06
     pant
    0.06
     lar
    0.06
     digging
    0.06
     باغ
    0.06
    (args
    0.06
     deployments
    0.06
    .Box
    0.06
    {}↵
    0.06
    Activations Density 0.050%

    No Known Activations