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. 84597
    Prev
    Next
    INDEX
    Explanations

    plot synopses

    np_max-act · gemini-2.0-flash

    The neuron is tuned to words expressing pursuit or intent—terms like “determined,” “claim,” “after,” “seeking,” 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
    theory
    -0.07
    converted
    -0.07
    veal
    -0.07
    	Logger
    -0.06
     categorie
    -0.06
    omes
    -0.06
    interaction
    -0.06
    (Properties
    -0.06
     illust
    -0.06
     Clinic
    -0.06
    POSITIVE LOGITS
    onomic
    0.07
     điều
    0.07
     geil
    0.06
    şehir
    0.06
     konk
    0.06
     ชนะ
    0.06
    ๑
    0.06
     ماك
    0.06
    ]=[
    0.06
    승
    0.06
    Activations Density 0.028%

    No Known Activations