Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 40406
    Prev
    Next
    INDEX
    Explanations

    XML elements/tags

    np_max-act · gemini-2.0-flash

    The neuron detects XML/HTML tag syntax, i.e. angle-bracketed element names and related “tag” tokens.

    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
    	dialog
    -0.07
    _FRIEND
    -0.07
     segregated
    -0.06
     OP
    -0.06
    (compare
    -0.06
     separator
    -0.06
    并不
    -0.06
    设计
    -0.06
    "↵↵↵↵
    -0.06
    (klass
    -0.06
    POSITIVE LOGITS
    해
    0.07
    �
    0.06
     вар
    0.06
     sentinel
    0.06
    SQLException
    0.06
    .strategy
    0.06
    urance
    0.06
    .AutoSize
    0.06
     Edinburgh
    0.06
     ΔE
    0.06
    Activations Density 0.019%

    No Known Activations