© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 5979
    Prev
    Next
    INDEX
    Explanations

    elements related to shared living spaces and community-oriented amenities

    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
    678
    +0.09
    0.4%
    196
    +0.09
    0.4%
    459
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.09
    0.10
    459
    +0.09
    0.01
    59
    +0.08
    0.04
    Negative Logits
    ãĥ©ãĥ³
    -0.80
     Rebell
    -0.78
    aclysm
    -0.76
     comet
    -0.76
    iliate
    -0.75
    GGGG
    -0.75
     landfall
    -0.72
    ozyg
    -0.70
    akedown
    -0.70
    wana
    -0.69
    POSITIVE LOGITS
     seating
    1.54
     rooms
    1.36
     bedrooms
    1.33
     desks
    1.31
     upstairs
    1.29
     downstairs
    1.29
     Rooms
    1.27
     chairs
    1.26
     floor
    1.25
     benches
    1.25
    Activations Density 3.072%

    No Known Activations