© 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. 55895
    Prev
    Next
    INDEX
    Explanations

    magic spells curses

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on narrative elements of a fairy‐tale story—especially magic‐related words like “spell,” “cast,” and other fantasy/magic tokens.

    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
    \":
    -0.07
    571
    -0.07
     визнача
    -0.07
    Sharp
    -0.07
    GN
    -0.06
    的时候
    -0.06
    اين
    -0.06
     Nora
    -0.06
    �
    -0.06
    clude
    -0.06
    POSITIVE LOGITS
    .Design
    0.07
    RegularExpression
    0.06
     hover
    0.06
     backup
    0.06
    .inst
    0.06
    (cors
    0.06
    .level
    0.06
     富
    0.06
    _mesh
    0.06
    备注
    0.06
    Activations Density 0.028%

    No Known Activations