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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Attention SAE Research Paper
    3. GPT2-Small
    4. Attention Out
    5. 5-ATT-KK
    6. 44234
    Prev
    Next
    INDEX
    Explanations

    references to crimes and incidents of violence, particularly those involving bloggers and government critics

    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
    0:0.01
    1:0.01
    2:0.08
    3:0.08
    4:0.25
    5:0.03
    6:0.08
    7:0.19
    8:0.02
    9:0.05
    10:0.05
    11:0.10
    Negative Logits
    inge
    -1.49
    icious
    -1.43
     existence
    -1.34
    season
    -1.29
    icipated
    -1.29
    imentary
    -1.28
     season
    -1.25
     vocabulary
    -1.25
    evidence
    -1.24
     inaugural
    -1.21
    POSITIVE LOGITS
     Cutter
    1.45
     dstg
    1.45
     srfAttach
    1.45
    ateurs
    1.44
     elders
    1.42
     Painter
    1.36
     holders
    1.35
     elites
    1.33
    bees
    1.32
     fishermen
    1.31
    Activations Density 0.001%

    No Known Activations