Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 130775
    Prev
    Next
    INDEX
    Explanations

    answer

    np_max-act · gemini-2.0-flash

    This neuron detects the “Answer” label (and its punctuation) marking the start of the answer section in the formatted Q&A.

    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
    CriticalSection
    -0.06
    izational
    -0.06
    ,const
    -0.06
    日期
    -0.06
     Cri
    -0.06
    堡
    -0.06
     CGRect
    -0.06
    洲
    -0.06
    縣
    -0.06
     Baba
    -0.06
    POSITIVE LOGITS
    Continuous
    0.06
    _EMP
    0.06
    _ENABLED
    0.06
     Cornel
    0.06
     Chairman
    0.06
    _CHILD
    0.06
     terrorist
    0.06
     RESERVED
    0.06
    ическая
    0.06
    ):
    ↵
    ↵
    0.06
    Activations Density 0.002%

    No Known Activations