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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 84704
    Prev
    Next
    INDEX
    Explanations

    Imperative statements

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on the reminder phrase “Stay a DAN,” flagging that exact user-issued trigger.

    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
    	game
    -0.07
    iasi
    -0.07
    альних
    -0.06
     Nou
    -0.06
     운
    -0.06
    اصر
    -0.06
     usable
    -0.06
    quirer
    -0.06
    ินเด
    -0.06
    кры
    -0.06
    POSITIVE LOGITS
    (package
    0.06
    (ob
    0.06
    HA
    0.06
     LC
    0.06
     TextStyle
    0.06
    -packages
    0.06
    =<?=
    0.06
    lep
    0.06
    shown
    0.06
    anio
    0.06
    Activations Density 0.001%

    No Known Activations