© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 5510
    Prev
    Next
    INDEX
    Explanations

    words and phrases associated with cheating or dishonest behavior

    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
    146
    +0.11
    0.5%
    220
    +0.10
    0.5%
    190
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    32
    +0.11
    0.02
    767
    +0.10
    0.01
    288
    +0.09
    0.02
    Negative Logits
    DragonMagazine
    -0.73
     Archdemon
    -0.72
    ä½ľ
    -0.66
    00200000
    -0.64
    nesota
    -0.64
     Pione
    -0.62
    weeney
    -0.61
    ingen
    -0.61
    rums
    -0.61
    ed
    -0.60
    POSITIVE LOGITS
    ulative
    0.78
    yre
    0.78
     sheet
    0.77
    cheat
    0.77
    chet
    0.76
    herer
    0.75
     cheat
    0.74
     cheating
    0.72
    sheet
    0.70
    hard
    0.68
    Activations Density 0.074%

    No Known Activations