© 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. 19-RESID-POST-AA
    6. 111894
    Prev
    Next
    INDEX
    Explanations

    travel

    np_max-act · gemini-2.0-flash

    This neuron detects unusually long or repeated-character tokens and document-boundary/extended-input tokens—i.e., long continuous strings or markers that indicate large input blocks.

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31

    special tokens and markers that indicate conversational structure, particularly turn boundaries and role transitions in chat-formatted dialogue.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jdhshshs138

    chat-conversation boundary markers and special formatting tokens (like start/end of turns and headers).

    oai_token-act-pair · gpt-5Triggered by @jdhshshs138
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_19/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.19.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    �
    -0.08
     garments
    -0.07
    掛
    -0.07
     Carson
    -0.07
     empty
    -0.07
    АН
    -0.07
     standing
    -0.06
    敌
    -0.06
    accom
    -0.06
     тех
    -0.06
    POSITIVE LOGITS
     där
    0.07
    .sym
    0.07
     CGRectMake
    0.06
    墙
    0.06
     أيضا
    0.06
    >();
    0.06
    categoryId
    0.06
     pageTitle
    0.06
     товарів
    0.06
    .setMessage
    0.06
    Activations Density 15.233%

    No Known Activations