© 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. 15-RESID-POST-AA
    11. 8508
    Prev
    Next
    INDEX
    Explanations

    United

    np_max-act · gemini-2.0-flash

    This neuron detects character/name placeholder tokens (tokens like "NAME_..."/entity identifiers) in the text.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     kotlinx
    -0.07
    OrderBy
    -0.07
     verb
    -0.07
     kısa
    -0.06
     Verb
    -0.06
    τηκε
    -0.06
     randint
    -0.06
    >User
    -0.06
     Booking
    -0.06
    (component
    -0.06
    POSITIVE LOGITS
     Cob
    0.06
     kích
    0.06
     Election
    0.06
    /star
    0.06
    ANO
    0.06
    Nobody
    0.06
    ULSE
    0.06
     ê
    0.06
    몬
    0.06
    ACS
    0.06
    Activations Density 0.106%

    No Known Activations