© 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. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 5854
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly parentheses

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    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
    430
    +0.10
    0.4%
    84
    +0.10
    0.4%
    504
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    95
    +0.10
    0.01
    668
    +0.10
    0.01
    142
    +0.09
    0.01
    Negative Logits
     disadvantage
    -0.71
     excellence
    -0.66
    eric
    -0.65
     Triumph
    -0.62
     disadvantages
    -0.62
     adversary
    -0.61
    cale
    -0.59
    lishes
    -0.59
     wisdom
    -0.57
     Zi
    -0.56
    POSITIVE LOGITS
    arak
    0.75
    ACC
    0.74
     UCHIJ
    0.69
    áµ
    0.68
    Topic
    0.67
    EMOTE
    0.67
    inventoryQuantity
    0.67
    udeb
    0.66
    istar
    0.65
    channelAvailability
    0.65
    Activations Density 0.038%

    No Known Activations