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. 8390
    Prev
    Next
    INDEX
    Explanations

    or

    np_max-act · gemini-2.0-flash

    This neuron detects the conjunction “or,” i.e. the token used to present alternatives.

    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
     рес
    -0.07
     grande
    -0.07
    Activ
    -0.07
    -connect
    -0.06
    -ups
    -0.06
    ंब
    -0.06
    olest
    -0.06
    ponents
    -0.06
     antivirus
    -0.06
    ulta
    -0.06
    POSITIVE LOGITS
    !!↵↵
    0.07
     Carnegie
    0.07
    lue
    0.06
    cycle
    0.06
    `.`
    0.06
    ,true
    0.06
    [key
    0.06
     specialize
    0.06
    GREE
    0.06
    	Vec
    0.06
    Activations Density 0.021%

    No Known Activations