© 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. 5-ATT-KK
    11. 20901
    Prev
    Next
    INDEX
    Explanations

    the word "don't" and phrases that convey negation or refusal

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.19
    1:0.10
    2:0.04
    3:0.06
    4:0.06
    5:0.19
    6:0.04
    7:0.01
    8:0.09
    9:0.10
    10:0.03
    11:0.03
    Negative Logits
     attm
    -1.75
     NEC
    -1.59
    ーン
    -1.51
     coron
    -1.44
     裏�
    -1.41
    Balt
    -1.37
     trove
    -1.34
     reperto
    -1.34
    ゴン
    -1.33
     constellation
    -1.33
    POSITIVE LOGITS
    igrate
    1.93
    ittle
    1.92
    uce
    1.91
    raise
    1.80
    erd
    1.80
     weep
    1.77
    ilt
    1.76
     Reduce
    1.75
    ounge
    1.73
    umph
    1.66
    Activations Density 0.007%

    No Known Activations