© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 42869
    Prev
    Next
    INDEX
    Explanations

    .NewRequest

    np_max-act · gemini-2.0-flash

    This neuron specifically picks up on instruction‐style tokens (like “fill,” “in,” “parts,” “blanks”) that indicate areas to be filled in or placeholders in the prompt.

    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
    unsafe
    -0.07
    TE
    -0.06
    DE
    -0.06
    ORA
    -0.06
     RTE
    -0.06
    ample
    -0.06
    ्वप
    -0.06
    _raw
    -0.06
    or
    -0.05
     ignor
    -0.05
    POSITIVE LOGITS
    .NewRequest
    0.10
     TForm
    0.07
     webpack
    0.07
     ولی
    0.07
    Γ
    0.07
    €
    0.07
    .titleLabel
    0.07
     destined
    0.07
     mở
    0.06
     leverage
    0.06
    Activations Density 0.000%

    No Known Activations