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. 5053
    Prev
    Next
    INDEX
    Explanations

    underscores, which often indicate technical or code-related 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
    479
    +0.14
    0.7%
    6
    +0.14
    0.7%
    739
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.14
    0.03
    107
    +0.14
    0.02
    442
    +0.13
    0.02
    Negative Logits
     pup
    -0.82
     plague
    -0.79
     rele
    -0.77
     whistle
    -0.75
     surviving
    -0.74
     pee
    -0.73
     plateau
    -0.73
    ifax
    -0.73
     pacing
    -0.71
     deduct
    -0.71
    POSITIVE LOGITS
    POL
    1.27
    WINDOWS
    1.20
    PER
    1.19
    OFF
    1.17
    SIZE
    1.16
    MAT
    1.14
    STER
    1.10
    Vert
    1.10
    PART
    1.09
    OWN
    1.09
    Activations Density 0.058%

    No Known Activations