© 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. 10367
    Prev
    Next
    INDEX
    Explanations

    words related to authority or positioning

    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
    0:0.07
    1:0.21
    2:0.03
    3:0.03
    4:0.04
    5:0.21
    6:0.06
    7:0.02
    8:0.06
    9:0.09
    10:0.09
    11:0.04
    Negative Logits
     concede
    -2.06
     uncond
    -1.94
     withdraw
    -1.81
     gain
    -1.80
     relinqu
    -1.76
     destro
    -1.75
     underestimate
    -1.75
     endanger
    -1.75
     escalate
    -1.73
     contend
    -1.72
    POSITIVE LOGITS
    ius
    2.42
    arius
    2.06
    aeus
    2.02
    idates
    2.00
     XII
    1.96
     Scrolls
    1.96
     Tome
    1.93
     XIV
    1.92
     resemblance
    1.85
     Chronicles
    1.84
    Activations Density 0.000%

    No Known Activations