© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 10-RES_POST_32K-OAI
    11. 27584
    Prev
    Next
    INDEX
    Explanations

    expressions related to the search or quest for something

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    362
    +0.11
    0.5%
    639
    +0.11
    0.5%
    164
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.11
    0.04
    604
    +0.11
    0.04
    580
    +0.10
    0.05
    Negative Logits
    ensis
    -0.75
    lua
    -0.67
    ember
    -0.67
    sectional
    -0.65
    ocy
    -0.63
    panic
    -0.63
    ismo
    -0.62
    oji
    -0.62
    section
    -0.60
    >]
    -0.59
    POSITIVE LOGITS
     ways
    1.05
     solutions
    0.98
     loopholes
    0.94
     replacements
    0.92
     suitable
    0.87
     effic
    0.87
     cures
    0.86
     weaknesses
    0.83
     alternatives
    0.82
     successors
    0.82
    Activations Density 0.478%

    No Known Activations