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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 110214
    Prev
    Next
    INDEX
    Explanations

    diverse text excerpts

    np_max-act · gemini-2.0-flash

    The neuron is tuned to fire on the special markup and role‐label tokens that delimit and identify system/user/assistant message boundaries.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    project management-related tasks and content.

    oai_token-act-pair · gpt-4o-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
     [{
    -0.07
    uintptr
    -0.06
    	redirect
    -0.06
    амп
    -0.06
     داشت
    -0.06
     DTO
    -0.06
     快
    -0.06
    _JOB
    -0.06
    Mailer
    -0.06
     Feast
    -0.06
    POSITIVE LOGITS
    /apis
    0.07
     точки
    0.07
    foreign
    0.07
    .Xml
    0.06
    (messages
    0.06
    coal
    0.06
     hypoth
    0.06
    خدام
    0.06
     Heritage
    0.06
    /sources
    0.06
    Activations Density 0.390%

    No Known Activations