© 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. 44509
    Prev
    Next
    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens inside string literals that name or identify modules/plugins/interfaces (e.g. plugin names, domain‐style IDs, architecture or API identifiers).

    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
    -management
    -0.08
     хочу
    -0.07
     documento
    -0.07
    ropy
    -0.07
    795
    -0.07
    getReference
    -0.07
    装置
    -0.07
     story
    -0.07
    ований
    -0.06
    	dir
    -0.06
    POSITIVE LOGITS
     fashionable
    0.07
     Atmospheric
    0.06
     NOT
    0.06
     запас
    0.06
    án
    0.06
     CUSTOM
    0.06
     friendships
    0.06
    NUMBER
    0.06
     Visible
    0.05
     ugly
    0.05
    Activations Density 0.019%

    No Known Activations