© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 78252
    Prev
    Next
    INDEX
    Explanations

    introduction

    np_max-act · gemini-2.0-flash

    This neuron detects the structural tokens that mark the start of the assistant’s reply (e.g. the “assistant” header and its surrounding delimiter tokens).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    descriptions of chemical companies and their operations.

    oai_token-act-pair · gpt-4o-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
     Overrides
    -0.07
    隆
    -0.07
     여기
    -0.06
    -design
    -0.06
    800
    -0.06
    .utf
    -0.06
    يا
    -0.06
     includ
    -0.06
     Caller
    -0.06
    定义
    -0.06
    POSITIVE LOGITS
     fortunes
    0.08
     Met
    0.07
     Herbert
    0.07
    community
    0.06
     wore
    0.06
    /terms
    0.06
     Μα
    0.06
    ,“
    0.06
    HELP
    0.06
     DM
    0.06
    Activations Density 0.019%

    No Known Activations