© 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. 5393
    Prev
    Next
    INDEX
    Explanations

    references to dwarfs and hydrogen

    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
    48
    +0.17
    0.8%
    288
    +0.12
    0.5%
    373
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.17
    0.02
    48
    +0.12
    0.01
    578
    +0.10
    0.01
    Negative Logits
    soDeliveryDate
    -0.71
    otle
    -0.68
    EA
    -0.67
    ombie
    -0.66
     Dragonbound
    -0.65
    lis
    -0.65
    Lens
    -0.64
    ipop
    -0.64
    visual
    -0.62
     Clip
    -0.62
    POSITIVE LOGITS
     Fortress
    1.01
    fs
    0.80
    ald
    0.80
     Dwar
    0.76
    stru
    0.76
    acters
    0.75
    alf
    0.70
    ¬¼
    0.70
    stein
    0.69
    craft
    0.69
    Activations Density 0.082%

    No Known Activations