© 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. 5462
    Prev
    Next
    INDEX
    Explanations

    phrases introducing examples or lists

    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
    133
    +0.10
    0.5%
    590
    +0.09
    0.4%
    459
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    625
    +0.10
    0.04
    498
    +0.09
    0.05
    71
    +0.09
    0.04
    Negative Logits
    undle
    -0.76
    uca
    -0.73
    endor
    -0.72
    ester
    -0.70
    ulhu
    -0.69
     Ended
    -0.69
    Boss
    -0.68
    byss
    -0.68
    etz
    -0.67
    urch
    -0.67
    POSITIVE LOGITS
     examples
    1.24
     reasons
    1.14
     excerpts
    1.02
     ways
    1.02
     highlights
    1.00
     salient
    1.00
     observations
    0.99
     illust
    0.94
     facts
    0.93
     considerations
    0.92
    Activations Density 0.300%

    No Known Activations