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

    Neuronpedia

    Jacobian LensNEW
    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. 57429
    Prev
    Next
    INDEX
    Explanations

    Code

    np_max-act · gemini-2.0-flash

    This neuron detects special control tokens marking the start of header or metadata sections in the transcript (e.g., “<|start_header_id|>”).

    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
    AFP
    -0.07
    تهم
    -0.07
    風
    -0.07
    silver
    -0.07
     loc
    -0.06
    events
    -0.06
     Pizza
    -0.06
    Equality
    -0.06
    Point
    -0.06
    ونية
    -0.06
    POSITIVE LOGITS
     assertEquals
    0.06
    .gson
    0.06
    翰
    0.06
    _FAIL
    0.06
    リア
    0.06
     Chancellor
    0.06
    elf
    0.06
    _apps
    0.06
    _HE
    0.06
     Gson
    0.06
    Activations Density 0.109%

    No Known Activations