© 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. 4784
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate the generation or production of something

    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
    108
    +0.11
    0.5%
    677
    +0.10
    0.5%
    557
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    677
    +0.11
    0.07
    608
    +0.10
    0.09
    335
    +0.09
    0.05
    Negative Logits
     aside
    -0.71
    bug
    -0.67
     opin
    -0.64
    awar
    -0.64
    soType
    -0.64
     stigma
    -0.64
    inventoryQuantity
    -0.63
     qualifier
    -0.62
    Availability
    -0.62
     advancement
    -0.62
    POSITIVE LOGITS
    qua
    0.86
    ãĤ©
    0.78
     usable
    0.77
    ohyd
    0.70
    aughtered
    0.68
    ļé
    0.68
    İĭ
    0.67
    raq
    0.66
    roxy
    0.66
     composite
    0.64
    Activations Density 1.550%

    No Known Activations