© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 3785
    Prev
    Next
    INDEX
    Explanations

    occurrences of punctuation and specific structural features within the text

    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
    222
    +0.11
    0.5%
    674
    +0.11
    0.5%
    680
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    25
    +0.11
    0.02
    476
    +0.11
    0.01
    158
    +0.10
    0.01
    Negative Logits
    Ħ¢
    -0.73
    falls
    -0.68
    loads
    -0.67
    adoes
    -0.66
    cook
    -0.64
     Annotations
    -0.62
     hires
    -0.61
    videos
    -0.61
    eers
    -0.60
    runs
    -0.60
    POSITIVE LOGITS
     guiActive
    0.80
     opposite
    0.79
    76561
    0.76
    essim
    0.74
     Indie
    0.72
     oppos
    0.70
    soType
    0.70
    onomous
    0.66
    Downloadha
    0.66
    tek
    0.66
    Activations Density 0.113%

    No Known Activations