© 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. 128300
    Prev
    Next
    INDEX
    Explanations

    Trying different solutions

    np_max-act · gemini-2.0-flash

    The neuron fires most strongly on code‐like or markup tokens (e.g. angle brackets, tags, class names), i.e. it detects snippets of code or HTML/templating syntax.

    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
    logradouro
    -0.07
    _COMPLETED
    -0.07
    apult
    -0.06
     jméno
    -0.06
    Compar
    -0.06
     기술
    -0.06
     βασ
    -0.06
     hecho
    -0.06
     consist
    -0.06
     Clear
    -0.06
    POSITIVE LOGITS
    centaje
    0.07
     روسی
    0.07
    fighters
    0.06
     ++
    0.06
    .getAbsolutePath
    0.06
     Aquarium
    0.06
     zdravot
    0.06
     mise
    0.06
     xml
    0.06
    _buttons
    0.06
    Activations Density 0.016%

    No Known Activations