© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 126265
    Prev
    Next
    INDEX
    Explanations

    Numbers 3 and 4

    np_max-act · gemini-2.0-flash

    This neuron detects superscripted numeric reference or author‐affiliation markers (e.g. the little “^1,” “^2,” etc., that label citations or author footnotes).

    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
    -hook
    -0.07
     Viewer
    -0.06
    ITHER
    -0.06
    (Rect
    -0.06
    hest
    -0.06
    ních
    -0.06
     Jehovah
    -0.06
    掛
    -0.06
     helmet
    -0.06
    ما
    -0.06
    POSITIVE LOGITS
     aim
    0.08
    σουν
    0.07
    utsch
    0.07
     Ž
    0.06
     Χ
    0.06
    TON
    0.06
     [{
    0.06
    multiply
    0.06
    監
    0.06
    이슈
    0.06
    Activations Density 0.003%

    No Known Activations