© 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. 5375
    Prev
    Next
    INDEX
    Explanations

    forms of communication and expressions of thoughts

    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
    756
    +0.14
    0.6%
    351
    +0.10
    0.5%
    373
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    720
    +0.14
    0.12
    351
    +0.10
    0.08
    393
    +0.10
    0.11
    Negative Logits
    cend
    -0.63
    ENCY
    -0.63
    coni
    -0.62
     arrivals
    -0.60
     Northern
    -0.58
     gunmen
    -0.58
     dual
    -0.58
    air
    -0.58
    ursday
    -0.57
     Frontier
    -0.57
    POSITIVE LOGITS
    peat
    0.76
    soType
    0.74
     cherish
    0.69
    isSpecialOrderable
    0.68
     share
    0.68
     cried
    0.65
    empt
    0.64
    aughed
    0.63
    osen
    0.62
     spoil
    0.61
    Activations Density 1.528%

    No Known Activations