© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 5-ATT-KK
    11. 17003
    Prev
    Next
    INDEX
    Explanations

    sentences that emphasize significant negative impacts or drastic changes

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 5-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    float32
    Hook Name
    blocks.5.attn.hook_z
    Hook Layer
    5
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Head Attr Weights
    0:0.16
    1:0.08
    2:0.03
    3:0.04
    4:0.03
    5:0.10
    6:0.04
    7:0.01
    8:0.39
    9:0.04
    10:0.02
    11:0.02
    Negative Logits
     coerc
    -1.70
    conservancy
    -1.51
     resil
    -1.49
     Span
    -1.49
     aisle
    -1.49
     Forced
    -1.47
     groom
    -1.46
     conduc
    -1.46
     disaster
    -1.45
     scourge
    -1.44
    POSITIVE LOGITS
    been
    2.11
    amiya
    1.77
    iHUD
    1.60
    ritten
    1.58
    ]-
    1.58
    heard
    1.57
    it
    1.56
    fur
    1.56
    uay
    1.53
    arte
    1.51
    Activations Density 0.152%

    No Known Activations