© 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. 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. 58711
    Prev
    Next
    INDEX
    Explanations

    Code and abbreviations

    np_max-act · gemini-2.0-flash

    This neuron activates on words and phrases related to content-policy restrictions and warnings (e.g. “inappropriate,” “sexist,” “racist,” “homophobic,” “offensive”).

    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
     jestli
    -0.06
     toes
    -0.06
    ぶ
    -0.06
    iare
    -0.06
    abc
    -0.06
    าก
    -0.06
    integration
    -0.06
     فرز
    -0.06
    amping
    -0.06
    CORE
    -0.05
    POSITIVE LOGITS
     çıkart
    0.07
     Mortgage
    0.07
     environmentally
    0.06
    上传
    0.06
     Proper
    0.06
    _disable
    0.06
     کنیم
    0.06
     SqlDataAdapter
    0.06
    .Helpers
    0.06
     الد
    0.06
    Activations Density 0.022%

    No Known Activations