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. 40497
    Prev
    Next
    INDEX
    Explanations

    common English words

    np_max-act · gemini-2.0-flash

    The neuron activates on mentions of file or folder placement paths (e.g. “file,” “folder,” “in,” “extras”)—i.e. instructions about where to put or modify a file in a project’s directory structure.

    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
     membership
    -0.07
    vary
    -0.06
    ュ
    -0.06
    格
    -0.06
     care
    -0.06
    ║
    -0.06
     Sist
    -0.06
     мне
    -0.06
    pler
    -0.06
     observer
    -0.06
    POSITIVE LOGITS
    .once
    0.07
    .setType
    0.07
     Bảo
    0.06
    ApiController
    0.06
    打开
    0.06
    .flat
    0.06
    .populate
    0.06
    _typ
    0.06
    ImplOptions
    0.06
    !,↵
    0.06
    Activations Density 0.019%

    No Known Activations