© 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
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 126951
    Prev
    Next
    INDEX
    Explanations

    agile

    np_max-act · gemini-2.0-flash

    The neuron fires on descriptive words emphasizing physical speed or nimbleness (e.g. “quick,” “agile,” “dodging”), signaling it’s detecting adjectives/adverbs that highlight rapid, graceful movement.

    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
    י�
    -0.06
    7
    -0.06
    629
    -0.06
    iterals
    -0.06
     movies
    -0.06
    595
    -0.06
    /km
    -0.06
    770
    -0.06
    wdx
    -0.06
     sélection
    -0.06
    POSITIVE LOGITS
     uphol
    0.07
     agility
    0.07
    фер
    0.07
     versatility
    0.07
    oping
    0.06
     πολυ
    0.06
     equip
    0.06
    نگی
    0.06
    pled
    0.06
    )animated
    0.06
    Activations Density 0.014%

    No Known Activations