© 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 EvalsExportsinterp-engineNEWAPI 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. 57989
    Prev
    Next
    INDEX
    Explanations

    Technical texts

    np_max-act · gemini-2.0-flash

    The neuron is primarily detecting tokens that appear in the system‐message header and instruction section (i.e. the “You are playing the role of…” metadata).

    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
     binder
    -0.06
     VO
    -0.06
    SKU
    -0.06
     vascular
    -0.06
    _rd
    -0.06
     kicker
    -0.06
     addictive
    -0.06
     hasattr
    -0.06
     flushing
    -0.06
     Loft
    -0.06
    POSITIVE LOGITS
     문
    0.07
    org
    0.06
     "",
↵
    0.06
    -E
    0.06
     Mỹ
    0.06
     вид
    0.06
    :{
↵
    0.06
    ras
    0.06
     Celebr
    0.06
    ในว
    0.06
    Activations Density 0.003%

    No Known Activations