© 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. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Attention Out - 32k
    9. 3-ATT_32K-OAI
    11. 3370
    Prev
    Next
    INDEX
    Explanations

    legal and formal language related to licensing and certification

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-att_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-attn-out-SAEs/v5_32k_layer_3
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.3.hook_attn_out
    Hook Layer
    3
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    448
    +0.11
    0.5%
    756
    +0.10
    0.5%
    548
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    479
    +0.11
    0.01
    456
    +0.10
    0.01
    175
    +0.10
    0.00
    Negative Logits
     checkpoints
    -0.73
    ÏĦ
    -0.72
     retaliate
    -0.71
     inflicting
    -0.68
     violently
    -0.68
    Ö¼
    -0.67
    á½
    -0.67
    cli
    -0.66
     invading
    -0.65
     topp
    -0.65
    POSITIVE LOGITS
    CLASSIFIED
    0.79
     LLP
    0.76
    âĹ¼
    0.76
    çīĪ
    0.76
     memoir
    0.74
    archment
    0.71
     Signed
    0.70
     vocational
    0.68
     Licensed
    0.68
     Professional
    0.67
    Activations Density 0.024%

    No Known Activations