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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 91121
    Prev
    Next
    INDEX
    Explanations

    tens

    np_max-act · gemini-2.0-flash

    This neuron detects words specifying digit positions (like “units,” “tens,” “hundreds,” “thousands,” “ten‐thousands,” “hundred‐thousands,” “millions,” etc.) in numeric‐digit questions.

    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
     woo
    -0.06
    庄
    -0.06
     olmadığı
    -0.06
     osp
    -0.06
     mandatory
    -0.06
     spiders
    -0.06
     трьох
    -0.06
    っても
    -0.06
    анных
    -0.06
    .SystemColors
    -0.06
    POSITIVE LOGITS
     масс
    0.08
    ۴۰
    0.07
     tens
    0.06
    Cause
    0.06
    40
    0.06
    $body
    0.06
    ltre
    0.06
     Accord
    0.06
    �
    0.06
    search
    0.06
    Activations Density 0.003%

    No Known Activations