© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 7074
    Prev
    Next
    INDEX
    Explanations

    phrases indicating contrast or opposition in arguments

    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
    482
    +0.10
    0.4%
    712
    +0.10
    0.4%
    683
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    720
    +0.10
    0.09
    351
    +0.10
    0.07
    285
    +0.09
    0.04
    Negative Logits
    word
    -0.65
    aurus
    -0.64
    appro
    -0.64
     [&
    -0.63
    CAP
    -0.59
     commence
    -0.57
    romancer
    -0.57
     vet
    -0.56
    rave
    -0.56
    leted
    -0.56
    POSITIVE LOGITS
    because
    1.44
     because
    1.43
     Because
    1.43
    Because
    1.36
    cause
    1.14
     due
    1.03
     bec
    0.99
     bc
    0.98
    due
    0.94
    ecause
    0.93
    Activations Density 1.741%

    No Known Activations