© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 2-ATT-KK
    11. 23251
    Prev
    Next
    INDEX
    Explanations

    phrases related to reflection on various topics such as past events, personal tendencies, or philosophical changes

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    phrases or clauses that indicate reflection or contemplation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    float32
    Hook Name
    blocks.2.attn.hook_z
    Hook Layer
    2
    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
    1:0.02
    2:0.13
    3:0.06
    4:0.30
    5:0.05
    6:0.02
    7:0.02
    8:0.17
    9:0.09
    10:0.05
    11:0.01
    Negative Logits
     checked
    -1.13
    oa
    -1.11
     respir
    -1.10
     Adren
    -1.10
     Mud
    -1.09
     parach
    -1.08
     shipped
    -1.07
     Route
    -1.06
     jump
    -1.06
     AAA
    -1.05
    POSITIVE LOGITS
    dayName
    1.58
    pter
    1.47
     Remastered
    1.40
    peak
    1.38
    pard
    1.37
    chio
    1.37
    orean
    1.34
    antz
    1.32
    ration
    1.32
    reality
    1.32
    Activations Density 0.011%

    No Known Activations