© 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. 39717
    Prev
    Next
    INDEX
    Explanations

    phrases related to evaluations or judgments about subjects or statements

    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
    0:0.01
    1:0.01
    2:0.09
    3:0.16
    4:0.14
    5:0.03
    6:0.04
    7:0.10
    8:0.09
    9:0.07
    10:0.12
    11:0.09
    Negative Logits
     sqor
    -1.68
    knit
    -1.37
    ourney
    -1.36
    +,
    -1.31
    TL
    -1.29
    oliath
    -1.18
     Unle
    -1.17
    ?),
    -1.13
    …)
    -1.13
    icrobial
    -1.13
    POSITIVE LOGITS
     ``
    1.50
     ----------------------------------------------------------------
    1.32
     admitting
    1.26
     conced
    1.24
     invalid
    1.23
     loopholes
    1.21
     accus
    1.18
    arat
    1.16
     sarc
    1.16
     defeat
    1.13
    Activations Density 0.024%

    No Known Activations