© 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. 13039
    Prev
    Next
    INDEX
    Explanations

    I

    np_max-act · gemini-2.0-flash

    The neuron signals when the assistant is issuing a first‐person refusal or apology (e.g. “I’m sorry,” “I cannot,” “I apologize”).

    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
    toa
    -0.08
    tearDown
    -0.07
    rıca
    -0.07
    .servers
    -0.07
     MySqlCommand
    -0.07
    InputLabel
    -0.07
    тя
    -0.06
    極
    -0.06
    RecyclerView
    -0.06
    rellas
    -0.06
    POSITIVE LOGITS
    IDDLE
    0.07
     immigration
    0.07
    /ic
    0.07
     Lesson
    0.06
    lev
    0.06
    -II
    0.06
    zoek
    0.06
    environment
    0.06
     callable
    0.06
     afirm
    0.06
    Activations Density 0.025%

    No Known Activations