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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 130574
    Prev
    Next
    INDEX
    Explanations

    rite

    np_max-act · gemini-2.0-flash

    The neuron specializes in spotting numbered step or procedure markers (e.g. “Step 1:”, “1.”, “2.”) in instructional text.

    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
    .ads
    -0.07
     pay
    -0.06
     Pay
    -0.06
    584
    -0.06
    Scan
    -0.06
    -type
    -0.06
    show
    -0.06
     shown
    -0.06
    (","
    -0.06
     FP
    -0.06
    POSITIVE LOGITS
     bou
    0.08
     prav
    0.08
    	document
    0.08
     गत
    0.07
    -INF
    0.07
     Blogger
    0.07
     khuôn
    0.07
    مت
    0.07
    �
    0.07
     comput
    0.07
    Activations Density 0.001%

    No Known Activations