© 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. 90787
    Prev
    Next
    INDEX
    Explanations

    questions and answers

    np_max-act · gemini-2.0-flash

    This neuron activates on instructional cue words telling the model to respond—particularly tokens in the phrase “and then answer the question.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    conversations that incorporate instructional language and pedagogical strategies.

    oai_token-act-pair · gpt-4o-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
     kost
    -0.06
    �이
    -0.06
     Indians
    -0.06
     الت
    -0.06
     �
    -0.06
    、三
    -0.06
     memoria
    -0.06
    Tuple
    -0.06
    させる
    -0.06
    พบ
    -0.06
    POSITIVE LOGITS
    \Blueprint
    0.07
    MapView
    0.07
     otev
    0.07
     dua
    0.07
    .LabelControl
    0.07
     beyaz
    0.06
    //$
    0.06
    .cf
    0.06
    арх
    0.06
    	add
    0.06
    Activations Density 0.033%

    No Known Activations