© 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
    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. 73909
    Prev
    Next
    INDEX
    Explanations

    command

    np_max-act · gemini-2.0-flash

    The neuron activates on comment lines that mention code‐generation metadata—e.g. “Code generated by the command…,” “THIS FILE IS GENERATED BY THE COMMAND…,” and related “do not edit” headings.

    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
    	source
    -0.07
    vis
    -0.07
    clicked
    -0.07
    án
    -0.07
    _use
    -0.07
     read
    -0.07
    دی
    -0.06
    (scan
    -0.06
    PASSWORD
    -0.06
    +:
    -0.06
    POSITIVE LOGITS
     prism
    0.06
     everlasting
    0.06
     tweets
    0.06
     DIFF
    0.06
     inhab
    0.06
    igated
    0.06
     staple
    0.06
     Asus
    0.06
    ’util
    0.06
    臺
    0.06
    Activations Density 0.003%

    No Known Activations