© 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. 4867
    Prev
    Next
    INDEX
    Explanations

    references to the Unicode character 'u'

    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
    324
    +0.13
    0.6%
    25
    +0.13
    0.6%
    534
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.13
    0.02
    238
    +0.13
    0.01
    670
    +0.11
    0.01
    Negative Logits
    GBT
    -0.81
    GB
    -0.75
    GY
    -0.75
    WN
    -0.74
    isphere
    -0.73
    RAY
    -0.71
    ILLE
    -0.70
    KER
    -0.68
    enhagen
    -0.67
    HT
    -0.65
    POSITIVE LOGITS
    selves
    0.76
    othes
    0.73
     CLSID
    0.71
    terday
    0.64
     sorts
    0.61
     sole
    0.59
    alon
    0.59
     unsc
    0.58
    username
    0.58
     mastermind
    0.58
    Activations Density 0.096%

    No Known Activations