© 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. 22958
    Prev
    Next
    INDEX
    Explanations

    phrases related to contrasting or contradictory statements

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

    discussions of paradoxes and contradictions in various contexts

    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
    0:0.05
    1:0.03
    2:0.15
    3:0.05
    4:0.18
    5:0.07
    6:0.03
    7:0.04
    8:0.06
    9:0.19
    10:0.07
    11:0.03
    Negative Logits
    untarily
    -1.38
    ebus
    -1.16
    unts
    -1.16
     refurb
    -1.10
    emouth
    -1.10
    ufact
    -1.08
     contrace
    -1.08
    SPA
    -1.07
     servicing
    -1.06
    esses
    -1.06
    POSITIVE LOGITS
     paradox
    1.56
    ּ
    1.35
     dich
    1.32
     causation
    1.32
     miscon
    1.30
    undrum
    1.28
     whereby
    1.26
     constructs
    1.25
    gravity
    1.25
     inconsistency
    1.23
    Activations Density 0.008%

    No Known Activations