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. 85901
    Prev
    Next
    INDEX
    Explanations

    religion

    np_max-act · gemini-2.0-flash

    This neuron detects archaic, verse‐style scriptural phrasing (e.g. words like “Verily,” “say,” “unto,” “proclaiming,” “Comforter”) that mark the beginnings or formal language of scripture passages.

    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
    Labor
    -0.07
     Alone
    -0.07
    useRalative
    -0.06
    桐
    -0.06
     왕
    -0.06
    AVED
    -0.06
    獲
    -0.06
    ルド
    -0.06
     वस
    -0.06
    adastrar
    -0.06
    POSITIVE LOGITS
    díl
    0.06
    mín
    0.06
     mein
    0.06
     embedded
    0.06
    opaque
    0.06
     resonance
    0.06
     Paragraph
    0.06
     doctrine
    0.06
    rezent
    0.06
     tiểu
    0.06
    Activations Density 0.005%

    No Known Activations