Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 55166
    Prev
    Next
    INDEX
    Explanations

    comma

    np_max-act · gemini-2.0-flash

    The neuron picks out direct‐speech warning or commanding words (imperative and deontic modals like “don’t,” “shouldn’t,” “warn,” “Suit yourself,” etc.).

    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
    -worker
    -0.08
    Slinky
    -0.07
    buah
    -0.06
     sized
    -0.06
     prayers
    -0.06
     refugee
    -0.06
     dicts
    -0.06
    .routing
    -0.06
     usage
    -0.06
    -free
    -0.06
    POSITIVE LOGITS
     hollow
    0.07
    countries
    0.06
    认
    0.06
    CHECK
    0.06
    ,本
    0.06
    Feb
    0.06
    �
    0.06
    ,看
    0.06
    ,一
    0.06
     indem
    0.06
    Activations Density 0.038%

    No Known Activations