© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4314
    Prev
    Next
    INDEX
    Explanations

    terms related to background checks and background information

    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
    575
    +0.12
    0.5%
    680
    +0.11
    0.5%
    422
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    641
    +0.12
    0.02
    575
    +0.11
    0.02
    225
    +0.10
    0.02
    Negative Logits
    nia
    -0.71
    nard
    -0.68
    oda
    -0.67
    enberg
    -0.67
    emon
    -0.66
    selves
    -0.65
    DOM
    -0.65
    ora
    -0.64
    esta
    -0.64
    ById
    -0.62
    POSITIVE LOGITS
     checks
    0.94
     check
    0.81
     briefings
    0.75
     Checks
    0.75
     information
    0.74
     knowledge
    0.73
     info
    0.71
    checks
    0.70
     investigations
    0.70
     considerations
    0.67
    Activations Density 0.132%

    No Known Activations