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. 104286
    Prev
    Next
    INDEX
    Explanations

    -

    np_max-act · gemini-2.0-flash

    The neuron consistently fires on the “Defendants” label in the case‐caption header lines of appellate opinions.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    references to legal case information and court proceedings.

    oai_token-act-pair · gpt-4o-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
     مرکز
    -0.08
    ็ม
    -0.07
     acknowledgement
    -0.07
    肩
    -0.06
     signatures
    -0.06
     داده
    -0.06
     ◆
    -0.06
     entitlement
    -0.06
     صند
    -0.06
    zd
    -0.06
    POSITIVE LOGITS
    (Function
    0.07
    ("("
    0.07
    acyj
    0.07
    {-
    0.06
    (not
    0.06
    vidia
    0.06
    acio
    0.06
     {/*
    0.06
    llvm
    0.06
    eters
    0.06
    Activations Density 0.003%

    No Known Activations