© 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. 92243
    Prev
    Next
    INDEX
    Explanations

    question-answering involving the user

    np_max-act · gemini-2.0-flash

    The neuron activates on second-person references—especially occurrences of “you” (and related “your,” “why you are,” etc.) when the text directly addresses the reader.

    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
    //----------------------------------------------------------------------------------------------------------------
    -0.06
     breakpoints
    -0.06
    .setTextSize
    -0.06
    Stack
    -0.06
     yourselves
    -0.06
    漂
    -0.06
    akespeare
    -0.06
     Shot
    -0.06
    UV
    -0.06
    肉
    -0.06
    POSITIVE LOGITS
    andidate
    0.07
    ;text
    0.07
     Việt
    0.07
    .Char
    0.06
    findById
    0.06
    ighest
    0.06
    ẫn
    0.06
    _elapsed
    0.06
    invert
    0.06
    (balance
    0.06
    Activations Density 0.081%

    No Known Activations