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. 130606
    Prev
    Next
    INDEX
    Explanations

    Dashes and parentheses

    np_max-act · gemini-2.0-flash

    This neuron fires on LaTeX fragments that express L¹‐convergence in measure theory—e.g. the integral “\int_X |f_n−f| dμ<ε,” L₁ norms, the function symbol f (and subscripts), and the backslash commands around them.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    terms and concepts related to mathematical convergence in sequences and functions.

    oai_token-act-pair · gpt-4o-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
    Asc
    -0.07
    delta
    -0.06
    Heap
    -0.06
    措施
    -0.06
    �
    -0.06
    spe
    -0.06
    tz
    -0.06
    ('-
    -0.06
    ่ต
    -0.06
    \Test
    -0.06
    POSITIVE LOGITS
     collaborated
    0.07
     klin
    0.07
     kiş
    0.07
     الم
    0.06
     saja
    0.06
     gang
    0.06
     giao
    0.06
     wiped
    0.06
     знаком
    0.06
     faced
    0.06
    Activations Density 0.005%

    No Known Activations