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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 84664
    Prev
    Next
    INDEX
    Explanations

    Companies and Production

    np_max-act · gemini-2.0-flash

    The neuron is strongly active on tokens in the assistant’s generated introduction paragraphs (i.e. text in the model’s reply rather than the user’s prompt).

    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
    出口
    -0.08
     Ю
    -0.07
     Skip
    -0.07
    Events
    -0.06
     Stark
    -0.06
     Barbar
    -0.06
    [H
    -0.06
    事件
    -0.06
    ゥ
    -0.06
     nhiệt
    -0.06
    POSITIVE LOGITS
     starter
    0.06
     aerospace
    0.06
    909
    0.06
    ่อน
    0.06
    kode
    0.06
    thetic
    0.06
    alaxy
    0.06
    Transpose
    0.06
    HELP
    0.06
    Evaluate
    0.06
    Activations Density 0.094%

    No Known Activations