© 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. 89905
    Prev
    Next
    INDEX
    Explanations

    complex systems

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens that reference chart-based questions and their step-by-step programmatic decomposition (e.g., “Question:”, “Column names”, “Program to answer the question by decomposing”, and other chart/question/program keywords).

    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
     besser
    -0.07
    -0.06
     Thiết
    -0.06
    (attributes
    -0.06
    am
    -0.06
     Harvard
    -0.06
    (js
    -0.06
    -0.06
     AppModule
    -0.06
    otide
    -0.06
    POSITIVE LOGITS
    ('+
    0.08
     замен
    0.07
     wear
    0.07
     vibrator
    0.07
     OF
    0.07
     arab
    0.06
    rou
    0.06
     dred
    0.06
     ساختمان
    0.06
     آبی
    0.06
    Activations Density 0.049%

    No Known Activations