© 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. 5420
    Prev
    Next
    INDEX
    Explanations

    negative phrases or concepts

    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
    373
    +0.15
    0.7%
    372
    +0.09
    0.4%
    691
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    152
    +0.15
    0.03
    290
    +0.09
    0.03
    691
    +0.09
    0.02
    Negative Logits
    merce
    -0.75
     wise
    -0.68
     locals
    -0.61
    mens
    -0.58
    æѦ
    -0.58
    pedia
    -0.57
     manpower
    -0.57
     privilege
    -0.56
     eh
    -0.56
    vt
    -0.56
    POSITIVE LOGITS
    door
    0.80
    office
    0.75
     Forth
    0.73
    opped
    0.72
     porch
    0.71
    arching
    0.71
    retch
    0.70
    quarters
    0.68
    up
    0.67
    row
    0.67
    Activations Density 0.082%

    No Known Activations