© 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. 73707
    Prev
    Next
    INDEX
    Explanations

    Instructions or usage

    np_max-act · gemini-2.0-flash

    This neuron specializes in detecting imperative or procedural instruction language—words like “using,” “use,” “try,” and similar terms that signal step-by-step guidance.

    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
     tự
    -0.07
    Wal
    -0.07
     vouchers
    -0.07
    fa
    -0.06
     begr
    -0.06
     replies
    -0.06
    Tank
    -0.06
    .DataTable
    -0.06
    Cons
    -0.06
     fund
    -0.06
    POSITIVE LOGITS
    ัวหน
    0.07
    =\"";↵
    0.07
     tüm
    0.07
    ernals
    0.06
    [--
    0.06
    ONY
    0.06
    τήσεις
    0.06
     aylık
    0.06
     sprink
    0.06
    WindowTitle
    0.06
    Activations Density 0.060%

    No Known Activations