© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 43330
    Prev
    Next
    INDEX
    Explanations

    :

    np_max-act · gemini-2.0-flash

    This neuron detects test‐harness command lines in comments—specifically the “RUN:” markers (the colon after RUN) and pipe separators in Clang/LLVM lit test directives.

    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
     Odin
    -0.06
     Bazı
    -0.06
    ึ
    -0.06
     Mädchen
    -0.06
    兹
    -0.06
    .geo
    -0.06
     testcase
    -0.05
    ps
    -0.05
     incontro
    -0.05
    ‌رس
    -0.05
    POSITIVE LOGITS
     raced
    0.08
    )");↵
    0.07
    (system
    0.07
     {})↵
    0.07
     configured
    0.07
     augmented
    0.07
     CREATE
    0.07
    )],↵
    0.07
    _GUID
    0.07
    .Multi
    0.07
    Activations Density 0.001%

    No Known Activations