© 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. 3-ATT-KK
    11. 14309
    Prev
    Next
    INDEX
    Explanations

    phrases related to the concept of turning off or shutting down

    oai_token-act-pair · gpt-3.5-turbo
    Triggered by @bot

    phrases involving the words "shut" and "out," along with related terms indicating isolation or withdrawal

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Head Attr Weights
    0:0.03
    1:0.02
    2:0.07
    3:0.32
    4:0.02
    5:0.02
    6:0.07
    7:0.08
    8:0.03
    9:0.12
    10:0.05
    11:0.12
    Negative Logits
     Liberties
    -1.45
    arcity
    -1.43
    iffs
    -1.37
    eport
    -1.29
     RELEASE
    -1.21
    apego
    -1.17
    itia
    -1.16
    aez
    -1.13
     Uriel
    -1.11
     Kou
    -1.10
    POSITIVE LOGITS
     schedule
    1.15
     desks
    1.14
     sec
    1.12
     bulb
    1.11
    hound
    1.08
     Firefly
    1.07
     streak
    1.07
     rout
    1.06
     mold
    1.05
     orderly
    1.04
    Activations Density 0.006%

    No Known Activations