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

    Neuronpedia

    APICircuit 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. 86189
    Prev
    Next
    INDEX
    Explanations

    Please

    np_max-act · gemini-2.0-flash

    The neuron mainly activates on polite request markers, specifically the token “Please.”

    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
    ifting
    -0.07
    cluster
    -0.07
     hatta
    -0.07
     wings
    -0.07
     selecting
    -0.06
          
    -0.06
    culator
    -0.06
    .setAttribute
    -0.06
    exter
    -0.06
     washer
    -0.06
    POSITIVE LOGITS
    ARN
    0.07
     」
    0.06
     don
    0.06
    WithEmailAndPassword
    0.06
     فران
    0.06
    彼
    0.06
    02
    0.06
    elix
    0.06
    να
    0.06
     експ
    0.06
    Activations Density 0.010%

    No Known Activations