© 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. 68545
    Prev
    Next
    INDEX
    Explanations

    Writing passages

    np_max-act · gemini-2.0-flash

    This neuron detects clause-introduction words or discourse markers (e.g. “that,” “about,” “que,” “dizendo,” “falando”) used to frame or specify the content of an instruction.

    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
     worn
    -0.08
    [k
    -0.07
     Cars
    -0.07
     Books
    -0.07
     Projects
    -0.07
     Duch
    -0.07
    部
    -0.07
     Rolex
    -0.07
     designs
    -0.07
     rehab
    -0.06
    POSITIVE LOGITS
     сохран
    0.07
    faculty
    0.06
    stantiate
    0.06
    �
    0.06
    lendirme
    0.06
    周收录
    0.06
     �
    0.06
     держ
    0.06
     gdy
    0.06
     Schema
    0.06
    Activations Density 0.072%

    No Known Activations