© 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. 4278
    Prev
    Next
    INDEX
    Explanations

    references to averages across various metrics

    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
    15
    +0.09
    0.4%
    408
    +0.09
    0.4%
    688
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    21
    +0.09
    0.04
    408
    +0.09
    0.04
    122
    +0.09
    0.03
    Negative Logits
    ornings
    -0.98
    ãĤ¤ãĥĪ
    -0.87
    ptives
    -0.84
    models
    -0.80
    events
    -0.79
    ceans
    -0.78
    PAC
    -0.75
    ãĤ¦ãĤ¹
    -0.75
    missions
    -0.74
     Cups
    -0.72
    POSITIVE LOGITS
     amount
    1.02
     footprint
    0.93
     number
    0.92
     swath
    0.88
     discrepancy
    0.86
     barrier
    0.85
     person
    0.84
     umbrella
    0.83
     recommendation
    0.82
     thing
    0.82
    Activations Density 0.175%

    No Known Activations