© 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. 4834
    Prev
    Next
    INDEX
    Explanations

    the word "that" and its variations in different contexts

    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
    655
    +0.10
    0.5%
    274
    +0.09
    0.4%
    662
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    274
    +0.10
    0.10
    662
    +0.09
    0.10
    598
    +0.09
    0.07
    Negative Logits
    leeve
    -0.65
    aukee
    -0.64
    ãĥĺ
    -0.60
    aunder
    -0.59
    omet
    -0.57
    IDE
    -0.57
     dstg
    -0.56
    pill
    -0.56
    pring
    -0.56
    omorphic
    -0.56
    POSITIVE LOGITS
     doesn
    1.12
     begs
    1.10
     means
    1.09
     isn
    1.06
     wasn
    1.05
     includes
    1.05
     contradicts
    1.05
     hasn
    1.04
     implies
    1.02
     sounds
    1.01
    Activations Density 0.256%

    No Known Activations