© 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
    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. 129842
    Prev
    Next
    INDEX
    Explanations

    instructions and requirements

    np_max-act · gemini-2.0-flash

    The neuron detects instructions or prompts that tell the model to follow or reference the system/user instructions (e.g. “you must complete the assistant answer based on the system instructions”).

    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
    ManyToOne
    -0.06
     russian
    -0.06
    subseteq
    -0.06
     csvfile
    -0.06
    дая
    -0.06
    材料
    -0.06
    char
    -0.06
     curing
    -0.06
     steep
    -0.06
    Cast
    -0.06
    POSITIVE LOGITS
    [b
    0.07
    ерим
    0.06
    ımı
    0.06
    NR
    0.06
    рою
    0.06
    овано
    0.06
    .gg
    0.06
    со
    0.06
     Researchers
    0.06
     registering
    0.06
    Activations Density 0.006%

    No Known Activations