© 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. 9-ATT-KK
    11. 8135
    Prev
    Next
    INDEX
    Explanations

    references to various films and their attributes, including their reception and context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 9-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.attn.hook_z
    Hook Layer
    9
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.15
    1:0.02
    2:0.08
    3:0.05
    4:0.04
    5:0.05
    6:0.08
    7:0.06
    8:0.28
    9:0.08
    10:0.04
    11:0.03
    Negative Logits
     Saskatchewan
    -4.28
     CJ
    -3.68
     Edmonton
    -3.53
    atchewan
    -3.52
    monton
    -3.44
     Texans
    -3.33
     Manitoba
    -3.29
     ML
    -3.22
    QB
    -3.21
     RG
    -3.20
    POSITIVE LOGITS
     Venice
    8.63
     Ven
    6.47
    Ven
    5.18
     Cannes
    4.27
     Leonardo
    4.03
     Marina
    4.00
     Naples
    3.95
     Italians
    3.85
     Ves
    3.85
     Aless
    3.74
    Activations Density 0.001%

    No Known Activations