© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 59418
    Prev
    Next
    INDEX
    Explanations

    to

    np_max-act · gemini-2.0-flash

    This neuron detects bracketed action descriptions or scene directions (e.g., the “[I sit on the couch…]” style narrative stage directions).

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
    	rm
    -0.06
    ycin
    -0.06
    不要
    -0.06
     reim
    -0.06
    issing
    -0.06
     genomes
    -0.06
    ุก
    -0.06
    .g
    -0.06
    udies
    -0.06
    Customers
    -0.06
    POSITIVE LOGITS
     ผล
    0.07
     Night
    0.07
     Май
    0.07
     :=↵
    0.06
     outliers
    0.06
     mommy
    0.06
     ณ
    0.06
    launch
    0.06
     perl
    0.06
    크
    0.06
    Activations Density 0.098%

    No Known Activations