© 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. 38368
    Prev
    Next
    INDEX
    Explanations

    calls to action for reporting crimes or suspicious activities

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    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
    0:0.02
    1:0.01
    2:0.07
    3:0.19
    4:0.24
    5:0.02
    6:0.03
    7:0.05
    8:0.05
    9:0.05
    10:0.11
    11:0.11
    Negative Logits
     disproportionately
    -2.02
     unwanted
    -1.76
     discriminated
    -1.74
     wrongly
    -1.73
     unequal
    -1.71
     escaping
    -1.70
     stuffing
    -1.69
     accumulating
    -1.68
     partying
    -1.68
     oppressed
    -1.67
    POSITIVE LOGITS
    HQ
    1.95
    ascript
    1.93
     Kamp
    1.87
    ゴン
    1.86
     Advis
    1.83
     Archives
    1.80
     Investigations
    1.79
    istries
    1.76
     Tours
    1.73
     Directions
    1.72
    Activations Density 0.009%

    No Known Activations