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

    Neuronpedia

    Jacobian LensNEW
    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. 56692
    Prev
    Next
    INDEX
    Explanations

    technical applications

    np_max-act · gemini-2.0-flash

    This neuron fires on tokens that appear in the assistant’s (model-generated) reply rather than the user’s input or metadata, effectively marking the assistant’s own output.

    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
    cmd
    -0.06
    argest
    -0.06
    aurus
    -0.06
     діяль
    -0.06
     Std
    -0.06
    茶
    -0.06
    FINAL
    -0.06
    .card
    -0.06
    apyrus
    -0.06
     lign
    -0.06
    POSITIVE LOGITS
     hooked
    0.06
    .pg
    0.06
    	initialize
    0.06
    �始化
    0.06
    .Top
    0.06
     shutting
    0.06
     Logging
    0.06
    .Path
    0.06
    stricted
    0.06
    sı
    0.06
    Activations Density 0.049%

    No Known Activations