© 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
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 57462
    Prev
    Next
    INDEX
    Explanations

    sports articles

    np_max-act · gemini-2.0-flash

    This neuron flags short headline‐style lead-in phrases (e.g. “Olivo and Figgins playing time,” “Lehner playing at a high level,” “Wright could return as early,” “Hurst Getting Closer,” “Kerley a 3rd down machine,” “Gray can’t find groove”).

    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
    classnames
    -0.08
     inning
    -0.07
     Swing
    -0.06
     Hidden
    -0.06
     Classical
    -0.06
     startling
    -0.06
    rowned
    -0.06
    CA
    -0.06
    	ds
    -0.06
    _pts
    -0.06
    POSITIVE LOGITS
     věc
    0.07
    people
    0.07
     …↵↵
    0.07
    líb
    0.06
     acquaint
    0.06
    ствие
    0.06
     기타
    0.06
    っても
    0.06
     位
    0.06
     사람들이
    0.06
    Activations Density 0.070%

    No Known Activations