© 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. 5340
    Prev
    Next
    INDEX
    Explanations

    specific numerical references and citation patterns in academic texts

    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
    +0.17
    0.8%
    0
    +0.10
    0.5%
    218
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.17
    0.06
    176
    +0.10
    0.03
    61
    +0.10
    0.04
    Negative Logits
    ãĥİ
    -0.77
    omorphic
    -0.69
     underestimated
    -0.66
     bottleneck
    -0.66
     cryptocurrencies
    -0.63
     dysph
    -0.61
     disrupted
    -0.61
     plummeted
    -0.60
     esc
    -0.60
     bonded
    -0.60
    POSITIVE LOGITS
     Ibid
    1.93
     ib
    1.29
     Cf
    1.20
     See
    0.95
     cf
    0.95
     Compare
    0.88
     supra
    0.81
    See
    0.81
     pp
    0.80
     *)
    0.79
    Activations Density 0.558%

    No Known Activations