© 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. 10-ATT-KK
    11. 12811
    Prev
    Next
    INDEX
    Explanations

    negations and the use of the word "to" in conjunction with verbs related to calling, fixing, and breaking

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_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.10.attn.hook_z
    Hook Layer
    10
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.04
    1:0.03
    2:0.12
    3:0.05
    4:0.06
    5:0.33
    6:0.03
    7:0.08
    8:0.03
    9:0.12
    10:0.02
    11:0.03
    Negative Logits
     because
    -3.21
    …)
    -2.89
    …]
    -2.81
    ★★
    -2.80
    because
    -2.78
    ––
    -2.78
    -->
    -2.74
    ……
    -2.73
    ...)
    -2.72
     ourselves
    -2.69
    POSITIVE LOGITS
     unspecified
    3.09
    mobi
    2.78
    idable
    2.68
     pas
    2.57
     parted
    2.50
     ye
    2.42
    solete
    2.30
     spar
    2.30
     0004
    2.27
     awa
    2.27
    Activations Density 0.002%

    No Known Activations