© 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. 8-RES_POST_32K-OAI
    11. 7100
    Prev
    Next
    INDEX
    Explanations

    common auxiliary verbs and their usage in sentences

    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
    218
    +0.14
    0.7%
    566
    +0.09
    0.4%
    437
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.14
    0.16
    55
    +0.09
    0.16
    566
    +0.09
    0.13
    Negative Logits
    agonist
    -0.78
    udos
    -0.75
    sequently
    -0.73
     oneself
    -0.71
    aughs
    -0.70
     Examination
    -0.70
    enegger
    -0.70
    Therefore
    -0.69
    ocard
    -0.69
     Flavoring
    -0.68
    POSITIVE LOGITS
     crappy
    1.12
     awesome
    1.11
     shitty
    1.09
     funky
    1.05
     nice
    1.03
     alright
    0.99
     kinda
    0.99
     okay
    0.98
     messed
    0.98
     awful
    0.97
    Activations Density 1.045%

    No Known Activations