© 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. 5332
    Prev
    Next
    INDEX
    Explanations

    references to the word "you" and its related context in conversations

    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
    55
    +0.17
    0.8%
    102
    +0.15
    0.7%
    512
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.17
    0.04
    566
    +0.15
    0.04
    674
    +0.11
    0.03
    Negative Logits
    iox
    -0.89
    regate
    -0.85
    ggles
    -0.78
    monary
    -0.78
    ulton
    -0.76
    ieve
    -0.75
    cue
    -0.75
    erald
    -0.74
    imize
    -0.72
    ossom
    -0.72
    POSITIVE LOGITS
     gonna
    1.12
     going
    1.08
     supposed
    1.05
     poised
    1.00
     indicative
    0.99
     compatible
    0.98
     kidding
    0.93
     synonymous
    0.93
     ready
    0.93
     incompatible
    0.91
    Activations Density 0.135%

    No Known Activations