© 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. 10-RES_POST_32K-OAI
    11. 29700
    Prev
    Next
    INDEX
    Explanations

    negative contractions like "isn't"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    248
    +0.10
    0.5%
    752
    +0.10
    0.5%
    348
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    442
    +0.10
    0.22
    430
    +0.10
    0.14
    36
    +0.10
    0.15
    Negative Logits
    wolves
    -0.99
    ials
    -0.77
    rollers
    -0.77
    abase
    -0.75
    kee
    -0.75
    hetti
    -0.73
    ihar
    -0.73
    banks
    -0.71
    asers
    -0.71
    cles
    -0.70
    POSITIVE LOGITS
     exploit
    0.76
     endeavour
    0.65
     inconvenience
    0.64
     assumption
    0.63
     TOD
    0.61
     experiment
    0.61
     time
    0.61
     something
    0.61
     implication
    0.60
     rect
    0.60
    Activations Density 0.823%

    No Known Activations