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

    Neuronpedia

    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. 5775
    Prev
    Next
    INDEX
    Explanations

    evidence of warnings and citations needed in a context

    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
    356
    +0.13
    0.6%
    613
    +0.11
    0.5%
    513
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    573
    +0.13
    0.01
    591
    +0.11
    0.01
    600
    +0.10
    0.01
    Negative Logits
    tsky
    -0.70
     Advertisement
    -0.65
    igslist
    -0.63
     temptation
    -0.63
     Heads
    -0.63
     Jackets
    -0.63
    inches
    -0.59
    rency
    -0.59
     Babe
    -0.58
     lure
    -0.58
    POSITIVE LOGITS
    é¾įå
    0.86
    ãĤ·ãĥ£
    0.79
    soType
    0.76
    æ©
    0.72
    microsoft
    0.69
    channelAvailability
    0.69
    ForgeModLoader
    0.69
    è¿
    0.67
    craft
    0.67
    APH
    0.65
    Activations Density 0.026%

    No Known Activations