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

    Neuronpedia

    Jacobian LensNEW
    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. 6751
    Prev
    Next
    INDEX
    Explanations

    patterns of consonant-vowel combinations

    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
    197
    +0.09
    0.4%
    734
    +0.09
    0.4%
    572
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    701
    +0.09
    0.04
    503
    +0.09
    0.03
    232
    +0.09
    0.04
    Negative Logits
     AVG
    -0.65
     refrain
    -0.64
     Elves
    -0.63
     fruits
    -0.63
     Bal
    -0.62
     crosses
    -0.62
     Cerberus
    -0.61
     buck
    -0.61
     Mori
    -0.61
     elves
    -0.60
    POSITIVE LOGITS
    eatures
    1.10
    ravel
    1.02
    ream
    1.00
    oop
    1.00
    ound
    0.99
    ormal
    0.96
    ogging
    0.94
    rame
    0.93
    ocument
    0.93
    ree
    0.91
    Activations Density 0.304%

    No Known Activations