© 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. 128497
    Prev
    Next
    INDEX
    Explanations

    magic

    np_max-act · gemini-2.0-flash

    This neuron detects use of metaphorical or “magic”-style descriptors (e.g. words like “magic,” “formula,” “tool”) that cast features or processes as if they were supernatural or wondrous.

    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
     asshole
    -0.06
    <usize
    -0.06
    HorizontalAlignment
    -0.06
    '''↵↵
    -0.06
     длин
    -0.06
    .lst
    -0.06
     rightly
    -0.06
    .class
    -0.06
     disagree
    -0.05
    POSITIVE LOGITS
     depos
    0.07
     Zip
    0.07
    stalk
    0.06
     flora
    0.06
    hands
    0.06
     méthode
    0.06
    wand
    0.06
     Netherlands
    0.06
     техні
    0.06
     venues
    0.06
    Activations Density 0.029%

    No Known Activations