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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 91912
    Prev
    Next
    INDEX
    Explanations

    programming errors

    np_max-act · gemini-2.0-flash

    The neuron is keyed to first‐person pronouns (e.g. “I,” “I’ve,” etc.), firing whenever the author refers to themselves.

    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
     Bench
    -0.06
    	Model
    -0.06
    (patient
    -0.06
    =new
    -0.06
    amet
    -0.06
     Feed
    -0.06
     unborn
    -0.06
    upport
    -0.06
    tains
    -0.06
     MACHINE
    -0.06
    POSITIVE LOGITS
    riott
    0.07
     metros
    0.07
     sé
    0.06
     Sony
    0.06
     <<<
    0.06
    ाँ
    0.06
    式会社
    0.06
    DX
    0.06
     فرزند
    0.06
     شركة
    0.06
    Activations Density 0.082%

    No Known Activations