© 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. 10-RES_POST_32K-OAI
    11. 29809
    Prev
    Next
    INDEX
    Explanations

    references to chemical compounds and their effects

    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
    373
    +0.19
    0.9%
    217
    +0.10
    0.5%
    293
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.19
    -0.03
    314
    +0.10
    -0.01
    284
    +0.10
    0.02
    Negative Logits
    [_
    -0.86
    REDACTED
    -0.85
    CLASSIFIED
    -0.84
    ascript
    -0.84
    vier
    -0.81
    acqu
    -0.79
    utm
    -0.78
    gnu
    -0.75
    xff
    -0.74
     ILCS
    -0.73
    POSITIVE LOGITS
     questioning
    0.62
     governing
    0.61
     negotiation
    0.60
     way
    0.59
     cautioned
    0.58
     question
    0.58
     molecule
    0.58
     tourism
    0.54
     negotiator
    0.54
     controlled
    0.54
    Activations Density 0.309%

    No Known Activations