© 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. 3312
    Prev
    Next
    INDEX
    Explanations

    legal terminology related to obligations and prohibitions

    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
    459
    +0.10
    0.5%
    425
    +0.10
    0.5%
    192
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.10
    0.03
    551
    +0.10
    0.03
    517
    +0.10
    0.03
    Negative Logits
     Notting
    -0.81
    anmar
    -0.72
     Hugo
    -0.71
     braking
    -0.67
    ertodd
    -0.66
     leaping
    -0.66
    rontal
    -0.63
     underest
    -0.63
     helic
    -0.63
     Fedora
    -0.62
    POSITIVE LOGITS
    mun
    0.83
    lead
    0.71
     determine
    0.69
    ences
    0.65
     Accessed
    0.65
    accompanied
    0.64
    RES
    0.64
    omet
    0.61
     obtain
    0.61
    VE
    0.61
    Activations Density 0.502%

    No Known Activations