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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 44370
    Prev
    Next
    INDEX
    Explanations

    Mixed text snippets

    np_max-act · gemini-2.0-flash

    The neuron consistently lights up on the ChatML “Channel”-style speaker markers and their punctuation (e.g. “Channel”, the channel number, the colon, the surrounding quotes), i.e. the tokens that denote speaker turns.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    details related to senior living inquiries and assistance needs.

    oai_token-act-pair · gpt-4o-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
     мам
    -0.06
    來
    -0.06
    editor
    -0.06
     szer
    -0.06
    _gift
    -0.06
     barber
    -0.06
     перевір
    -0.06
     glam
    -0.06
     mixed
    -0.06
    ano
    -0.06
    POSITIVE LOGITS
     безопасности
    0.08
     оз
    0.07
     Institutions
    0.07
     Aut
    0.06
    Engineering
    0.06
     nouvelles
    0.06
    Holy
    0.06
     Поч
    0.06
    Aggregate
    0.06
     bt
    0.06
    Activations Density 0.002%

    No Known Activations