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. 84712
    Prev
    Next
    INDEX
    Explanations

    numbers

    np_max-act · gemini-2.0-flash

    The neuron primarily detects numeric tokens—digits, decimal numbers, currency amounts, and percentage values.

    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
    IALIZ
    -0.07
    	hr
    -0.06
     bik
    -0.06
     chees
    -0.06
     furry
    -0.06
     cmds
    -0.06
     FA
    -0.06
     frank
    -0.06
    (notification
    -0.06
     Castillo
    -0.06
    POSITIVE LOGITS
     ActionType
    0.07
    _True
    0.06
     Cir
    0.06
    -anchor
    0.06
     mData
    0.06
     правиль
    0.06
    .Calendar
    0.06
    عادة
    0.06
     만족
    0.06
    岛
    0.06
    Activations Density 0.011%

    No Known Activations