© 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. 13583
    Prev
    Next
    INDEX
    Explanations

    references to people, particularly in the context of reputation or societal role

    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
    0:0.03
    1:0.03
    2:0.05
    3:0.08
    4:0.19
    5:0.15
    6:0.05
    7:0.05
    8:0.09
    9:0.14
    10:0.02
    11:0.08
    Negative Logits
    ..."
    -2.49
    .?
    -2.45
    Redditor
    -2.44
    ,"
    -2.42
    Experts
    -2.37
     totaling
    -2.31
    ""
    -2.28
     Cosponsors
    -2.27
    !!"
    -2.27
    packages
    -2.26
    POSITIVE LOGITS
    ogue
    2.79
     foe
    2.32
     novelty
    2.28
     enemy
    2.21
     disliked
    2.20
    viation
    2.20
    leground
    2.13
     abbre
    2.13
     insurgency
    2.07
    umbn
    2.06
    Activations Density 0.003%

    No Known Activations