© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 12715
    Prev
    Next
    INDEX
    Explanations

    article snippets

    np_max-act · gemini-2.0-flash

    The neuron is picking up tokens that occur at the start of the assistant’s reply (e.g. “Hello,” “Girls,” or other greeting tokens), effectively marking the assistant’s generated-response segment.

    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
    せる
    -0.06
     eiusmod
    -0.06
    ех
    -0.06
    cli
    -0.06
     typingsSlinky
    -0.06
    .atomic
    -0.06
    Copy
    -0.06
     Advanced
    -0.06
    -states
    -0.06
    чик
    -0.06
    POSITIVE LOGITS
    гляд
    0.08
     القي
    0.07
    ="<?
    0.06
    :@{
    0.06
     blindness
    0.06
     accr
    0.06
     Pens
    0.06
     Çin
    0.06
     cbo
    0.06
    @g
    0.06
    Activations Density 0.004%

    No Known Activations