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. 1104
    Prev
    Next
    INDEX
    Explanations

    provide

    np_max-act · gemini-2.0-flash

    The neuron fires on modal and conditional verbs (e.g. “may,” “will,” “only,” “if,” “required”) that introduce policy rules or obligations.

    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
    -contained
    -0.06
    μή
    -0.06
     swollen
    -0.06
     candidates
    -0.06
     gala
    -0.06
     tiger
    -0.06
     caf
    -0.06
    ما
    -0.06
     dogs
    -0.06
    вед
    -0.06
    POSITIVE LOGITS
    ')?></
    0.08
     ^↵
    0.07
     GWei
    0.07
    bios
    0.06
    ($('<
    0.06
     Freel
    0.06
     Britt
    0.06
    	reg
    0.06
     Butterfly
    0.06
     Phil
    0.06
    Activations Density 0.011%

    No Known Activations