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. 40658
    Prev
    Next
    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron fires on colon characters used as key–value separators in structured data (e.g. the “users:” in “users: abc123 …”).

    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
    ंय
    -0.07
    цы
    -0.07
    <Option
    -0.07
    /py
    -0.07
     контроль
    -0.07
    bridge
    -0.06
    OfFile
    -0.06
    ","","
    -0.06
     affects
    -0.06
    。(
    -0.06
    POSITIVE LOGITS
    FAILED
    0.08
    异常
    0.07
     تست
    0.07
    uspend
    0.06
    fusc
    0.06
    кот
    0.06
    ordan
    0.06
    MAP
    0.06
     begging
    0.06
    emics
    0.06
    Activations Density 0.407%

    No Known Activations