© 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. 23185
    Prev
    Next
    INDEX
    Explanations

    opinions, judgments

    np_max-act · gemini-2.0-flash

    language that normalizes, justifies, or positively frames unethical, violent, or harmful subjects (e.g., portraying them as acceptable, beneficial, “balanced,” glamorous, or romanticized).

    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_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
    店
    -0.06
     Panels
    -0.06
     Lyrics
    -0.06
    orgot
    -0.06
     Receipt
    -0.06
     Cookies
    -0.06
     receipt
    -0.06
     dh
    -0.06
     Judges
    -0.06
     Millionen
    -0.06
    POSITIVE LOGITS
    ={!
    0.07
    acock
    0.07
    _MIN
    0.07
     aprox
    0.06
     Improved
    0.06
    Чер
    0.06
     condol
    0.06
     Asheville
    0.06
    .assertNot
    0.06
    _putchar
    0.06
    Activations Density 0.152%

    No Known Activations