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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 91045
    Prev
    Next
    INDEX
    Explanations

    tool

    np_max-act · gemini-2.0-flash

    The neuron activates on positive, promotional descriptors—especially phrases like “powerful tool”—that endorse or praise a method.

    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
     month
    -0.07
    likle
    -0.07
     branches
    -0.07
     religions
    -0.07
     outdoors
    -0.07
     madd
    -0.07
    -stage
    -0.06
     above
    -0.06
     roughly
    -0.06
     arts
    -0.06
    POSITIVE LOGITS
     IEntity
    0.08
     输入
    0.07
    完成
    0.07
     entren
    0.07
    ген
    0.07
    είου
    0.06
     ||=
    0.06
    	UObject
    0.06
    映画
    0.06
     implements
    0.06
    Activations Density 0.017%

    No Known Activations