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. 86123
    Prev
    Next
    INDEX
    Explanations

    punctuation

    np_max-act · gemini-2.0-flash

    The neuron fires on the special header/metadata tokens (e.g. “<|start_header_id|>”, step-by-step markers) that delimit or label sections of the assistant’s response.

    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
    érience
    -0.08
    .ByteString
    -0.07
     stigma
    -0.07
     kok
    -0.06
     brig
    -0.06
    quake
    -0.06
     суду
    -0.06
    (Channel
    -0.06
     by
    -0.06
    .completed
    -0.06
    POSITIVE LOGITS
     =>↵
    0.07
    _Rect
    0.07
    _SE
    0.06
    videos
    0.06
    	link
    0.06
    РА
    0.06
     toilet
    0.06
     الانت
    0.06
     )}↵
    0.06
     lol
    0.06
    Activations Density 0.207%

    No Known Activations