© 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. 29897
    Prev
    Next
    INDEX
    Explanations

    Chat

    np_max-act · gemini-2.0-flash

    The neuron activates on the word “chat” (and its variants like “chatting” or “chat rooms”).

    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
    ระเบ
    -0.07
     olanak
    -0.07
     Operand
    -0.07
    =P
    -0.07
     мона
    -0.06
    .Unique
    -0.06
     الکترون
    -0.06
     Knee
    -0.06
     Parkway
    -0.06
     Roots
    -0.06
    POSITIVE LOGITS
     chat
    0.14
     Chat
    0.13
    Chat
    0.11
    chat
    0.10
    _chat
    0.09
     chatting
    0.09
    (Chat
    0.09
    (chat
    0.08
    .Chat
    0.08
     bit
    0.08
    Activations Density 0.024%

    No Known Activations