© 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. 13070
    Prev
    Next
    INDEX
    Explanations

    news agencies or sources

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

    references to news agencies or media sources

    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
    0:0.03
    1:0.05
    2:0.17
    3:0.07
    4:0.04
    5:0.03
    6:0.10
    7:0.06
    8:0.11
    9:0.12
    10:0.12
    11:0.06
    Negative Logits
     selves
    -0.97
    ription
    -0.94
    覚醒
    -0.85
     cle
    -0.83
     dile
    -0.83
     macros
    -0.83
     stains
    -0.82
     Ram
    -0.81
     differe
    -0.81
     detail
    -0.79
    POSITIVE LOGITS
     largeDownload
    1.09
    orthern
    1.08
    ミ
    1.06
     GOODMAN
    1.04
    %)
    1.01
     Jagu
    1.00
    ヤ
    0.99
    )",
    0.97
    usterity
    0.95
    )]
    0.94
    Activations Density 0.018%

    No Known Activations