© 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. 15-RESID-POST-AA
    11. 55757
    Prev
    Next
    INDEX
    Explanations

    questions/problems

    np_max-act · gemini-2.0-flash

    It detects first-person self-reference—places where the author says “I” or uses closely connected auxiliary/contracted forms to describe their own actions or experience.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     acquire
    -0.07
    DAO
    -0.06
     bio
    -0.06
    skými
    -0.06
    >$
    -0.06
    ση
    -0.06
     클래스
    -0.06
     С
    -0.06
    	DEBUG
    -0.06
    ßerdem
    -0.06
    POSITIVE LOGITS
     号
    0.07
    _sub
    0.07
     Perc
    0.06
    ائية
    0.06
    DBus
    0.06
    ạng
    0.06
    ับค
    0.06
    ीव
    0.06
    fol
    0.06
    _CHANGE
    0.06
    Activations Density 0.034%

    No Known Activations