© 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. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 9112
    Prev
    Next
    INDEX
    Explanations

    parts

    np_max-act · gemini-2.0-flash

    The neuron activates on instruction and organizational cue words—terms like “include,” “part,” “end,” “components,” and similar—marking places where the user is specifying how to structure or organize content.

    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
    ЙЙ
    -0.07
     Contributors
    -0.07
    .Matchers
    -0.06
    DAO
    -0.06
    NSURLSession
    -0.06
    eptal
    -0.06
     Yug
    -0.06
    ↵	
    ↵
    -0.06
    _curr
    -0.06
     Vu
    -0.06
    POSITIVE LOGITS
     obchod
    0.07
     erkek
    0.07
    يث
    0.07
    olithic
    0.06
    _kelas
    0.06
    becca
    0.06
    リカ
    0.06
    Modifiers
    0.06
     thats
    0.06
    .getLongitude
    0.06
    Activations Density 0.154%

    No Known Activations