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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 42046
    Prev
    Next
    INDEX
    Explanations

    10

    np_max-act · gemini-2.0-flash

    The neuron fires on occurrences of the “Windows 10” version name (i.e. the token “10” when it’s part of the Windows 10 mention).

    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
    |()↵
    -0.07
    itez
    -0.07
    brook
    -0.07
    亚洲
    -0.07
    Hashtable
    -0.06
     Todo
    -0.06
     Setup
    -0.06
    Buttons
    -0.06
    ([]*
    -0.06
     Negro
    -0.06
    POSITIVE LOGITS
    10
    0.07
    _remain
    0.06
     inspir
    0.06
     �
    0.06
     threading
    0.06
     sn
    0.06
    _ELEMENTS
    0.06
     accents
    0.06
     gains
    0.06
    incre
    0.06
    Activations Density 0.003%

    No Known Activations