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

    Neuronpedia

    Jacobian LensNEW
    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. 38128
    Prev
    Next
    INDEX
    Explanations

    stunts and tricks

    np_max-act · gemini-2.0-flash

    The neuron activates on descriptions of high-energy stunts and acrobatic tricks (e.g., flips, mid-air maneuvers) performed by stunt teams or extreme sports performers.

    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
    -metal
    -0.07
     administrative
    -0.07
     insistence
    -0.06
    _AGENT
    -0.06
     Ber
    -0.06
     thread
    -0.06
    アル
    -0.06
     Bar
    -0.06
    Ash
    -0.06
    Ber
    -0.06
    POSITIVE LOGITS
     выс
    0.06
    .th
    0.06
     urlencode
    0.06
    urances
    0.06
    .every
    0.06
     ανά
    0.06
    /");↵
    0.06
     соверш
    0.06
    ــ
    0.06
    raci
    0.06
    Activations Density 0.009%

    No Known Activations