© 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 EvalsExportsinterp-engineNEWAPI 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. 32009
    Prev
    Next
    INDEX
    Explanations

    financial metrics and discussions related to income and employment

    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
    232
    +0.21
    1.1%
    308
    +0.17
    0.9%
    430
    +0.17
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.21
    0.27
    145
    +0.17
    0.16
    176
    +0.17
    0.07
    Negative Logits
     senate
    -0.89
     elves
    -0.88
     aven
    -0.87
     queen
    -0.86
     estab
    -0.85
     satell
    -0.85
     skelet
    -0.85
     queens
    -0.83
     rampant
    -0.82
     regress
    -0.82
    POSITIVE LOGITS
     Needs
    1.66
     Coverage
    1.63
     Detail
    1.63
     Relationship
    1.60
     Problems
    1.59
     Accounts
    1.59
     Costs
    1.59
     Relations
    1.58
     Treatment
    1.57
     Process
    1.57
    Activations Density 6.953%

    No Known Activations