Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 5026
    Prev
    Next
    INDEX
    Explanations

    sequences of colons and special characters often found in coding or technical syntax

    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
    385
    +0.11
    0.5%
    289
    +0.11
    0.5%
    709
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    107
    +0.11
    0.01
    326
    +0.11
    0.01
    385
    +0.09
    0.01
    Negative Logits
     enriched
    -0.73
    ulkan
    -0.72
    iments
    -0.71
     drained
    -0.68
    brance
    -0.68
    hetti
    -0.66
    ously
    -0.64
    lar
    -0.64
     fres
    -0.64
     riddled
    -0.63
    POSITIVE LOGITS
    operator
    0.90
    NL
    0.89
    before
    0.82
    ãĥĵ
    0.81
    wcsstore
    0.79
    small
    0.79
    temp
    0.79
    âĸijâĸij
    0.79
    Delta
    0.77
    âĹı
    0.75
    Activations Density 0.058%

    No Known Activations