© 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. 41045
    Prev
    Next
    INDEX
    Explanations

    last

    np_max-act · gemini-2.0-flash

    The neuron is sensitive to modal/advisory language—words and phrases that offer warnings, recommendations, or qualify advice (e.g. “should,” “only,” “last resort,” “extreme,” “irreversible”).

    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
     merges
    -0.07
     自
    -0.07
    .newInstance
    -0.06
    čník
    -0.06
    了
    -0.06
     runnable
    -0.06
     FETCH
    -0.06
    .roles
    -0.06
     unsettling
    -0.06
     towards
    -0.06
    POSITIVE LOGITS
     ong
    0.07
    량
    0.07
    	card
    0.06
    �
    0.06
     خرد
    0.06
    Reviews
    0.06
    藥
    0.06
    복
    0.06
     Indi
    0.06
     директор
    0.06
    Activations Density 0.008%

    No Known Activations