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. 85863
    Prev
    Next
    INDEX
    Explanations

    Sports game summaries

    np_max-act · gemini-2.0-flash

    The neuron consistently flags play‐by‐play football commentary language—especially descriptions of defensive actions, player positions, and in‐game movements.

    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
     referees
    -0.06
    ASF
    -0.06
     лак
    -0.06
    inions
    -0.06
     stressing
    -0.06
     cdr
    -0.06
     basal
    -0.06
    폐
    -0.06
     governmental
    -0.06
     Bach
    -0.06
    POSITIVE LOGITS
    	pr
    0.06
    _scroll
    0.06
    (It
    0.06
    $view
    0.06
     NSArray
    0.06
    RUN
    0.06
    ivement
    0.06
    _cs
    0.06
    Paid
    0.06
    vyk
    0.06
    Activations Density 0.015%

    No Known Activations