© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 12647
    Prev
    Next
    INDEX
    Explanations

    Inability

    np_max-act · gemini-2.0-flash

    This neuron detects tokens at the start of error messages—especially negative-modal verbs like “Could,” “Cannot,” and “Unable” in “Could not…,” “Cannot…,” or “Unable to…” statements.

    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
     Ser
    -0.07
     responder
    -0.06
     debugging
    -0.06
     Bernard
    -0.06
     Kate
    -0.06
    transparent
    -0.06
    اسر
    -0.06
    itzer
    -0.06
     classifiers
    -0.06
    .Par
    -0.06
    POSITIVE LOGITS
     Present
    0.07
     level
    0.06
    /?
    0.06
    생
    0.06
    ?",
    0.06
    /control
    0.06
     çık
    0.06
    　
    0.06
     cush
    0.06
    造成
    0.06
    Activations Density 0.019%

    No Known Activations