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. 130198
    Prev
    Next
    INDEX
    Explanations

    sorry

    np_max-act · gemini-2.0-flash

    expressions of regret or apologies.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    This neuron detects expressions of apology (e.g., “I’m sorry”).

    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
     Force
    -0.07
    phere
    -0.06
    有的
    -0.06
    екси
    -0.06
     ####
    -0.06
     muff
    -0.06
    (Build
    -0.06
     relegated
    -0.06
    qe
    -0.06
     dẫn
    -0.06
    POSITIVE LOGITS
     sorry
    0.08
     sympath
    0.08
     tempted
    0.07
    opped
    0.07
    .syn
    0.07
    lander
    0.07
     Sorry
    0.07
    になり
    0.07
    keiten
    0.06
    aci
    0.06
    Activations Density 0.022%

    No Known Activations