© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 2-ATT-KK
    11. 22857
    Prev
    Next
    INDEX
    Explanations

    modes of staring or gazing in various settings

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

    instances of people staring or looking at something

    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
    0:0.07
    1:0.02
    2:0.14
    3:0.07
    4:0.10
    5:0.07
    6:0.02
    7:0.03
    8:0.20
    9:0.15
    10:0.05
    11:0.02
    Negative Logits
    cially
    -1.35
     redistributed
    -1.28
    soever
    -1.22
     Presbyter
    -1.21
     certified
    -1.21
     Associ
    -1.20
     trustee
    -1.16
     facult
    -1.15
     registered
    -1.12
     depend
    -1.08
    POSITIVE LOGITS
    ��
    1.85
    TextColor
    1.39
     blinking
    1.36
    lights
    1.30
    advertisement
    1.28
     horrors
    1.26
    livion
    1.25
    enance
    1.24
     helpless
    1.22
    reality
    1.19
    Activations Density 0.013%

    No Known Activations