© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 3-ATT-KK
    11. 13077
    Prev
    Next
    INDEX
    Explanations

    references to platforms where content is uploaded or shared

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    instances of user-generated content and uploads

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    float32
    Hook Name
    blocks.3.attn.hook_z
    Hook Layer
    3
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.02
    1:0.02
    2:0.08
    3:0.17
    4:0.02
    5:0.05
    6:0.06
    7:0.09
    8:0.08
    9:0.22
    10:0.07
    11:0.07
    Negative Logits
     actionGroup
    -1.52
    uncture
    -1.39
    terness
    -1.30
    armac
    -1.28
    dding
    -1.21
    natureconservancy
    -1.20
    Haunted
    -1.19
    leground
    -1.18
    adish
    -1.18
    ォ
    -1.16
    POSITIVE LOGITS
     spreadsheet
    1.25
     planner
    1.14
     PowerPoint
    1.14
     Lucas
    1.13
     Pastebin
    1.12
     onto
    1.10
     Garmin
    1.09
    abase
    1.08
     Tup
    1.07
     archive
    1.05
    Activations Density 0.005%

    No Known Activations