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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 26721
    Prev
    Next
    INDEX
    Explanations

    e

    np_max-act · gemini-2.0-flash

    error messages related to Python code execution.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    This neuron predominantly activates on exception‐handling constructs, specifically the exception variable (e) bound in an “except Exception as e” clause.

    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
     metres
    -0.07
     responds
    -0.07
     presented
    -0.07
    /info
    -0.07
    (alpha
    -0.06
    scores
    -0.06
     intim
    -0.06
     всіх
    -0.06
     hüc
    -0.06
     tease
    -0.06
    POSITIVE LOGITS
    塞
    0.06
    ัพท
    0.06
    .transform
    0.06
     Shiite
    0.06
    ited
    0.06
     �
    0.06
    <ActionResult
    0.06
     formulario
    0.06
    ετ
    0.06
    대
    0.06
    Activations Density 0.004%

    No Known Activations