© 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. 89762
    Prev
    Next
    INDEX
    Explanations

    lately

    np_max-act · gemini-2.0-flash

    The neuron detects words expressing personal physical or emotional discomfort (e.g. “heavy,” “lately,” “stress”).

    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
     vocals
    -0.06
    اوی
    -0.06
     FileInfo
    -0.06
    truth
    -0.06
     Bad
    -0.06
     Abram
    -0.06
    Destroy
    -0.06
     ΑΝ
    -0.06
    pid
    -0.06
     OPTIONS
    -0.06
    POSITIVE LOGITS
     kys
    0.07
     configure
    0.07
     meydana
    0.07
    细
    0.07
     lately
    0.06
     assez
    0.06
    _IL
    0.06
    	lbl
    0.06
     skip
    0.06
     laboratory
    0.06
    Activations Density 0.015%

    No Known Activations