Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 130342
    Prev
    Next
    INDEX
    Explanations

    question answering

    np_max-act · gemini-2.0-flash

    This neuron detects directive or meta-instruction passages—i.e. the parts of the text that tell the assistant how to behave.

    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
    orc
    -0.07
    585
    -0.07
     beet
    -0.07
    otation
    -0.07
     exporter
    -0.07
    415
    -0.07
     capacities
    -0.06
     drink
    -0.06
    .find
    -0.06
     capacity
    -0.06
    POSITIVE LOGITS
    allen
    0.07
     примі
    0.06
    ві
    0.06
    etzt
    0.06
    }$/
    0.06
     جنوبی
    0.06
    ічні
    0.06
    ickým
    0.06
    جة
    0.06
     getTime
    0.06
    Activations Density 0.205%

    No Known Activations