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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 4291
    Prev
    Next
    INDEX
    Explanations

    references to evaluations and assessments of companies or products

    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
    567
    +0.10
    0.5%
    396
    +0.10
    0.4%
    578
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    654
    +0.10
    0.07
    686
    +0.10
    0.05
    60
    +0.09
    0.07
    Negative Logits
    imar
    -0.92
    quished
    -0.68
    iots
    -0.67
    illery
    -0.67
    icken
    -0.67
    amon
    -0.65
    anus
    -0.65
    krit
    -0.64
    ortality
    -0.61
    arations
    -0.60
    POSITIVE LOGITS
     fit
    0.95
     best
    0.91
     suitable
    0.88
     weakest
    0.84
     strengths
    0.82
     nearest
    0.82
     optimal
    0.81
     maxim
    0.79
     suited
    0.79
     compatible
    0.79
    Activations Density 2.977%

    No Known Activations