© 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. 6045
    Prev
    Next
    INDEX
    Explanations

    negative numerical values or symbols

    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
    373
    +0.22
    1.0%
    427
    +0.10
    0.5%
    545
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.22
    0.01
    284
    +0.10
    0.01
    365
    +0.09
    0.01
    Negative Logits
    ription
    -0.83
    ufact
    -0.77
    ulhu
    -0.76
    ascript
    -0.75
    DEN
    -0.74
    initions
    -0.73
    ãĥ¼ãĥĨãĤ£
    -0.71
    SHARE
    -0.70
    insula
    -0.69
    ibaba
    -0.69
    POSITIVE LOGITS
    ier
    0.62
     elf
    0.60
    games
    0.58
    orns
    0.55
     shots
    0.55
     Demons
    0.55
     enclave
    0.52
    Games
    0.52
    izons
    0.52
    bies
    0.52
    Activations Density 0.060%

    No Known Activations