© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 29675
    Prev
    Next
    INDEX
    Explanations

    negation/opposites

    np_max-act · gemini-2.0-flash

    This neuron detects control tokens that mark message and speaker boundaries (e.g. end‐of‐header and end‐of‐text tags) in the conversation.

    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
     облад
    -0.06
     framerate
    -0.06
    much
    -0.06
     captivating
    -0.06
    opot
    -0.06
    еса
    -0.06
    worked
    -0.06
    ."↵↵↵
    -0.06
     enticing
    -0.06
     Jungle
    -0.06
    POSITIVE LOGITS
     Iterator
    0.07
     Update
    0.07
    (log
    0.07
     счет
    0.07
     EXP
    0.06
     smartphones
    0.06
    _POLL
    0.06
     modifiers
    0.06
    _make
    0.06
     Gallup
    0.06
    Activations Density 0.013%

    No Known Activations