© 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. 41589
    Prev
    Next
    INDEX
    Explanations

    Unexpected events

    np_max-act · gemini-2.0-flash

    The neuron detects mentions of unexpectedness or surprise, such as “surprise,” “caught … off guard,” and similar expressions.

    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
     Fotos
    -0.07
    frica
    -0.06
     lump
    -0.06
    紹
    -0.06
     sın
    -0.06
    acle
    -0.06
    \Dependency
    -0.06
    Dig
    -0.06
    íme
    -0.06
     dohod
    -0.06
    POSITIVE LOGITS
     surprised
    0.07
     unaware
    0.07
     surprise
    0.07
     revelations
    0.07
    tabla
    0.07
     tension
    0.06
    ندا
    0.06
    _Style
    0.06
    ины
    0.06
    инку
    0.06
    Activations Density 0.021%

    No Known Activations