© 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. 91221
    Prev
    Next
    INDEX
    Explanations

    collaborations

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens describing the model’s provenance—i.e. mentions of its development or joint training by institutions (like “developed,” “trained,” institution names, and dates).

    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
     Haupt
    -0.07
     هفت
    -0.07
    arring
    -0.06
    paring
    -0.06
     Bearings
    -0.06
     пен
    -0.06
     HID
    -0.06
    wash
    -0.06
     kort
    -0.06
    andest
    -0.06
    POSITIVE LOGITS
    (fs
    0.08
    <li
    0.06
    '])?
    0.06
    acea
    0.06
     nostalg
    0.06
    .Protocol
    0.06
    >>>(
    0.06
    341
    0.06
     acompañ
    0.06
     missed
    0.06
    Activations Density 0.004%

    No Known Activations