© 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. 40860
    Prev
    Next
    INDEX
    Explanations

    phrases related to consistency and permanence

    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
    2:0.11
    3:0.07
    4:0.24
    5:0.04
    6:0.05
    7:0.13
    8:0.10
    9:0.05
    10:0.05
    11:0.06
    Negative Logits
     millenn
    -1.49
     unden
    -1.40
     newfound
    -1.39
     induction
    -1.36
     reluct
    -1.30
    attribute
    -1.28
     ingenuity
    -1.26
     misunder
    -1.25
    ベ
    -1.21
    invoke
    -1.20
    POSITIVE LOGITS
    /-
    1.36
    utral
    1.34
     �
    1.33
    hest
    1.31
    lymp
    1.30
    ghazi
    1.29
    independent
    1.28
    Bir
    1.28
    orthy
    1.28
     Maine
    1.28
    Activations Density 0.001%

    No Known Activations