Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Attention SAE Research Paper
    3. GPT2-Small
    4. Attention Out
    5. 2-ATT-KK
    6. 22365
    Prev
    Next
    INDEX
    Explanations

    items related to economic or business news

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

    information related to trade deals, political events, and public opinions on social issues in various locations

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

    specific entities or organizations along with their characteristics or statistics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    float32
    Hook Name
    blocks.2.attn.hook_z
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.06
    1:0.03
    2:0.08
    3:0.22
    4:0.03
    5:0.09
    6:0.07
    7:0.05
    8:0.08
    9:0.06
    10:0.11
    11:0.06
    Negative Logits
     channels
    -1.22
    ernels
    -1.18
     establishments
    -1.17
     formations
    -1.16
     cabinets
    -1.14
     calendars
    -1.13
    urrencies
    -1.11
    inas
    -1.09
    doms
    -1.09
     prizes
    -1.08
    POSITIVE LOGITS
    ccording
    1.50
     spokesperson
    1.22
     ensued
    1.17
     staffer
    1.15
     difference
    1.13
     Difference
    1.12
     spokeswoman
    1.11
     example
    1.11
     spokesman
    1.07
     glance
    1.05
    Activations Density 0.103%

    No Known Activations