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. 86095
    Prev
    Next
    INDEX
    Explanations

    honest

    np_max-act · gemini-2.0-flash

    The neuron activates on occurrences of the word “honest” (including its translations) in the assistant’s self-descriptive statements.

    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
    .take
    -0.08
    ‌پ
    -0.08
     fık
    -0.07
     ^{↵
    -0.07
     ویرایش
    -0.07
    データ
    -0.06
    登録
    -0.06
    Scripts
    -0.06
     {\↵
    -0.06
    标题
    -0.06
    POSITIVE LOGITS
     Guide
    0.06
    ский
    0.06
     PSP
    0.06
    rawl
    0.06
    GROUP
    0.06
    'name
    0.06
     LOWER
    0.06
     STACK
    0.06
     gateway
    0.06
     stripper
    0.06
    Activations Density 0.003%

    No Known Activations