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. 85992
    Prev
    Next
    INDEX
    Explanations

    's

    np_max-act · gemini-2.0-flash

    This neuron fires on tokens that start new sections or sentences—i.e. the first, often capitalized, word beginning a heading or sentence.

    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
    ']>;↵
    -0.07
    _jwt
    -0.06
     Сем
    -0.06
     Bellev
    -0.06
     آموز
    -0.06
    —from
    -0.06
     інтер
    -0.06
     Desde
    -0.06
     Meat
    -0.06
    磁
    -0.06
    POSITIVE LOGITS
    ="-
    0.07
    ่ท
    0.06
    istrator
    0.06
    structured
    0.06
    ่อส
    0.06
     relatively
    0.06
    phe
    0.06
    sole
    0.06
    瑟
    0.06
    IT
    0.06
    Activations Density 0.037%

    No Known Activations