© 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. 15-RESID-POST-AA
    6. 101460
    Prev
    Next
    INDEX
    Explanations

    array

    np_max-act · gemini-2.0-flash

    prompts and responses involving explicit sexual roleplay setups, coercive in-character instructions, and erotic dialogue contexts.

    oai_token-act-pair · gpt-5Triggered by @vetterc0

    requests or exchanges that try to initiate explicit, sexualized roleplay with the assistant, often including coercive or boundary-pushing instructions.

    oai_token-act-pair · gpt-5Triggered 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
    ครอง
    -0.07
     Burb
    -0.07
     shimmer
    -0.06
     twee
    -0.06
     Pace
    -0.06
     Ginny
    -0.06
    แชม
    -0.06
    .parameters
    -0.06
     hodin
    -0.06
     retrieved
    -0.06
    POSITIVE LOGITS
    :UIAlert
    0.07
    所以
    0.07
    EDIT
    0.06
    ैद
    0.06
    StartupScript
    0.06
     resistance
    0.06
    :set
    0.06
    athroom
    0.06
    Persist
    0.06
     endure
    0.06
    Activations Density 0.040%

    No Known Activations