© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 57366
    Prev
    Next
    INDEX
    Explanations

    Decisions

    np_max-act · gemini-2.0-flash

    discussions about physical attributes and body image preferences.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron is highly sensitive to mid‐sized content words—especially persuasive, action‐oriented verbs and terms used in sales pitches (like “project,” “second,” “look,” “giving,” “share,” “listen,” “explore”)—marking language that urges or guides the listener toward next steps.

    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
    폰
    -0.07
     manera
    -0.07
    器
    -0.06
    仁
    -0.06
    략
    -0.06
     Radar
    -0.06
    児
    -0.06
    ели
    -0.06
     paragraphs
    -0.06
    anning
    -0.06
    POSITIVE LOGITS
    otyping
    0.06
    *g
    0.06
     самостоятель
    0.06
     списка
    0.06
     выпол
    0.06
    คว
    0.06
    .border
    0.06
    .Accessible
    0.06
    nodes
    0.06
     ส
    0.05
    Activations Density 0.198%

    No Known Activations