© 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
    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. 38369
    Prev
    Next
    INDEX
    Explanations

    They

    np_max-act · gemini-2.0-flash

    content related to explicit sexual themes.

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

    The neuron detects words that describe characters’ reactions or responses (e.g. watched, nodded, listened, taking).

    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
    診
    -0.07
     Docs
    -0.06
     акку
    -0.06
    Reddit
    -0.06
    _eye
    -0.06
    ddit
    -0.06
    kového
    -0.06
    pt
    -0.06
    자
    -0.06
    atego
    -0.06
    POSITIVE LOGITS
    .addHandler
    0.07
    British
    0.07
    /helper
    0.07
    French
    0.07
     mashed
    0.06
     Spanish
    0.06
    谋
    0.06
    UITableViewCell
    0.06
    [axis
    0.06
    .Spring
    0.06
    Activations Density 0.049%

    No Known Activations