© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 57061
    Prev
    Next
    INDEX
    Explanations

    legal liability

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on legal phrasing that signals vicarious liability (e.g. “vicariously liable for”), effectively detecting references to the respondeat-superior doctrine.

    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
     ruku
    -0.07
    ुध
    -0.06
    LANG
    -0.06
    wen
    -0.06
     }*/↵↵
    -0.06
    ToLocal
    -0.06
     mrt
    -0.06
    maxlength
    -0.06
    啊啊
    -0.06
    olang
    -0.06
    POSITIVE LOGITS
    coverage
    0.07
     Sacr
    0.06
     Sailor
    0.06
     IonicPage
    0.06
     pumpkin
    0.06
    (memory
    0.06
    0.06
     luggage
    0.06
     quil
    0.06
    -coordinate
    0.06
    Activations Density 0.003%

    No Known Activations