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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · GPT-OSS BatchTopK SAEs
    3. GPT-OSS-20B
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 78636
    Prev
    Next
    INDEX
    Explanations

    .exports

    np_max-act · gemini-2.0-flash

    has not been explained as it seems like the conclusion to an instruction script and does not pertain to any attention head behavior in the initial context given. If you need further clarification or details regarding attention head behaviors in neural networks, please feel free to ask!

    oai_attention-head · gpt-4oTriggered by @mlfinancebook
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-gpt-oss-20b/resid_post_layer_11/trainer_0
    Dataset (Dashboard)
    Various
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    _air
    -0.08
    Misc
    -0.07
    Hom
    -0.07
     azure
    -0.07
     efficiencies
    -0.07
     ach
    -0.07
    almost
    -0.07
    indlu
    -0.07
     admission
    -0.07
    اده
    -0.07
    POSITIVE LOGITS
     apostles
    0.08
     patri
    0.08
     Kathmandu
    0.08
     Clement
    0.08
     nursing
    0.08
     Ordin
    0.08
    läge
    0.08
     cultivating
    0.08
     Jem
    0.08
     contribut
    0.07
    Activations Density 0.001%

    No Known Activations