© 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. 41463
    Prev
    Next
    INDEX
    Explanations

    references to sporting events, particularly cricket and baseball, specifically in relation to World Cups and Series

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

    No Comments

    Head Attr Weights
    0:0.02
    1:0.01
    2:0.13
    3:0.05
    4:0.07
    5:0.04
    6:0.14
    7:0.03
    8:0.06
    9:0.03
    10:0.17
    11:0.20
    Negative Logits
    ALSE
    -1.57
    )=(
    -1.56
    abre
    -1.53
    enta
    -1.53
    stan
    -1.46
    SourceFile
    -1.40
    ipedia
    -1.40
    iste
    -1.38
    kef
    -1.36
    rament
    -1.35
    POSITIVE LOGITS
     backlog
    1.32
     Authorization
    1.24
     burgl
    1.24
     wrapper
    1.23
     revolving
    1.21
     crossover
    1.20
     leaks
    1.20
     entitlement
    1.19
     anonymously
    1.18
     remake
    1.14
    Activations Density 0.003%

    No Known Activations