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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 54942
    Prev
    Next
    INDEX
    Explanations

    blog posts/articles

    np_max-act · gemini-2.0-flash

    The neuron primarily detects the special token marking the end of a header block (the `<|end_header_id|>` marker).

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

    special chat-format tokens that mark the end of a role header, especially the boundary before an assistant message.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
    ニ
    -0.07
    ETO
    -0.07
     Tubes
    -0.06
    ollah
    -0.06
     Keith
    -0.06
    te
    -0.06
     cream
    -0.06
    _EXIST
    -0.06
    esen
    -0.06
     Login
    -0.06
    POSITIVE LOGITS
    (recv
    0.06
     스타
    0.06
     onClose
    0.06
    (da
    0.06
     >
    ↵
    0.06
     �
    0.06
     перева
    0.06
    +',
    0.06
    +'\
    0.06
     recount
    0.06
    Activations Density 0.102%

    No Known Activations