© 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. 92350
    Prev
    Next
    INDEX
    Explanations

    comma

    np_max-act · gemini-2.0-flash

    names of people, particularly notable figures or characters in various contexts.

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

    The neuron spikes on capitalized tokens—especially single‐letter or acronym‐style bits (e.g. “S” in Sábado, “E” in E.I., “K” in Kansai, “B” in BODIPY)—i.e. it’s looking for uppercase/name‐like fragments.

    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
     CultureInfo
    -0.08
    ($__
    -0.07
    الت
    -0.06
    のに
    -0.06
    raní
    -0.06
    ileen
    -0.06
    Psi
    -0.06
     bilingual
    -0.06
     shields
    -0.06
    .for
    -0.06
    POSITIVE LOGITS
    .uuid
    0.07
     pic
    0.07
     Вал
    0.06
     babe
    0.06
    ValueHandling
    0.06
     vypad
    0.06
     поля
    0.06
     histories
    0.06
    ٤
    0.06
     �
    0.06
    Activations Density 0.042%

    No Known Activations