Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Attention SAE Research Paper
    3. GPT2-Small
    4. Attention Out
    5. 2-ATT-KK
    6. 22530
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal or criminal procedures

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

    imperative and directive phrases

    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
    0:0.07
    1:0.02
    2:0.09
    3:0.12
    4:0.17
    5:0.06
    6:0.03
    7:0.02
    8:0.14
    9:0.15
    10:0.07
    11:0.03
    Negative Logits
     engulf
    -1.20
    interstitial
    -1.11
     stabilized
    -1.09
     desperately
    -1.07
     reinstated
    -1.05
    profits
    -1.04
    egal
    -1.04
     collapses
    -1.03
     collapsing
    -1.03
     Appears
    -1.02
    POSITIVE LOGITS
     Crim
    1.29
     beforehand
    1.26
     Consent
    1.25
     "{
    1.24
     Enough
    1.19
    esson
    1.17
     Scots
    1.15
    urses
    1.13
    USE
    1.12
    ���
    1.12
    Activations Density 0.040%

    No Known Activations