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

    Neuronpedia

    Jacobian LensNEW
    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. 30369
    Prev
    Next
    INDEX
    Explanations

    references to images or illustrations related to events

    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
    731
    +0.15
    0.7%
    281
    +0.11
    0.5%
    20
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    442
    +0.15
    0.05
    430
    +0.11
    0.05
    373
    +0.10
    -0.02
    Negative Logits
     poisoning
    -0.72
     disadvant
    -0.65
    ersen
    -0.65
    knife
    -0.64
     assailant
    -0.64
    anooga
    -0.64
     poisoned
    -0.63
     poison
    -0.63
     equilibrium
    -0.62
    cffff
    -0.62
    POSITIVE LOGITS
    omi
    0.75
    August
    0.71
    DAY
    0.70
    ĵ
    0.69
    um
    0.68
    isodes
    0.68
    am
    0.68
    )=(
    0.68
    ules
    0.67
    greg
    0.67
    Activations Density 0.477%

    No Known Activations