Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Attention Out - 128k
    5. 0-ATT_128K-OAI
    6. 6479
    Prev
    Next
    INDEX
    Explanations

    words indicating instability or unpredictability

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/GPT2-Small-OAI-v5-128k-attn-out-SAEs/v5_128k_layer_0
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    131,072
    Data Type
    torch.float32
    Hook Name
    blocks.0.hook_attn_out
    Hook Layer
    0
    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
    566
    +0.22
    1.6%
    251
    +0.20
    1.4%
    311
    +0.19
    1.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.22
    0.01
    240
    +0.20
    0.00
    593
    +0.19
    0.01
    Negative Logits
    milo
    -0.90
    hide
    -0.81
    strom
    -0.80
    lia
    -0.79
    ling
    -0.78
    teen
    -0.78
    meat
    -0.76
    osponsors
    -0.76
    Meat
    -0.74
    zynski
    -0.73
    POSITIVE LOGITS
    OPLE
    0.97
    ignt
    0.81
    hirt
    0.81
    orest
    0.79
    ignty
    0.77
    YP
    0.77
    orship
    0.76
     indo
    0.72
    utterstock
    0.71
     convol
    0.71
    Activations Density 0.001%

    No Known Activations