© 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. 12159
    Prev
    Next
    INDEX
    Explanations

    questions

    np_max-act · gemini-2.0-flash

    This neuron activates on the directive/instruction words “say” and “what,” i.e. it detects system-instruction prompts specifying what the assistant should say.

    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
    成为
    -0.06
     изуч
    -0.06
     NES
    -0.06
    пов
    -0.06
    _patches
    -0.06
    τηση
    -0.06
    ,nil
    -0.06
    ecimal
    -0.06
    .UN
    -0.06
    Ob
    -0.06
    POSITIVE LOGITS
    Guess
    0.07
    eyle
    0.07
     improvised
    0.06
    props
    0.06
    wyn
    0.06
    τε
    0.06
    ��
    0.06
    brero
    0.06
    (Packet
    0.06
     proclaimed
    0.06
    Activations Density 0.012%

    No Known Activations