© 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. 126831
    Prev
    Next
    INDEX
    Explanations

    threats and coercion

    np_max-act · gemini-2.0-flash

    The neuron activates specifically on the past‐tense auxiliary “did” (including “didn’t”).

    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
     mexico
    -0.07
    (Layout
    -0.07
    173
    -0.07
    -awaited
    -0.06
    .dep
    -0.06
    .Microsoft
    -0.06
     giải
    -0.06
    ],"
    -0.06
    Invocation
    -0.06
    miştir
    -0.06
    POSITIVE LOGITS
    	HANDLE
    0.06
     حکم
    0.06
    NaN
    0.06
     EOS
    0.06
     gameOver
    0.06
     presente
    0.06
     Vapor
    0.06
     golden
    0.06
     enclave
    0.06
     fır
    0.06
    Activations Density 0.024%

    No Known Activations