© 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. 4451
    Prev
    Next
    INDEX
    Explanations

    negations or words that indicate refusal

    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
    +0.13
    0.6%
    365
    +0.09
    0.4%
    546
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    212
    +0.13
    0.02
    696
    +0.09
    0.03
    697
    +0.09
    0.02
    Negative Logits
    senal
    -0.67
     Peng
    -0.65
     helicop
    -0.65
     Kirin
    -0.63
    Introduced
    -0.62
    ansas
    -0.59
     Powered
    -0.59
    SOURCE
    -0.59
     Genie
    -0.58
    gap
    -0.57
    POSITIVE LOGITS
    't
    1.11
    aught
    0.84
    aturally
    0.84
    ither
    0.82
    ately
    0.78
    oot
    0.77
    '>
    0.73
    atell
    0.72
    esides
    0.69
    ÃŃ
    0.69
    Activations Density 0.168%

    No Known Activations