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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 13542
    Prev
    Next
    INDEX
    Explanations

    mentions of the platform Flickr and its usage in online content sharing

    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
    311
    +0.22
    1.4%
    508
    +0.19
    1.3%
    578
    +0.17
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    524
    +0.22
    0.03
    334
    +0.19
    0.02
    294
    +0.17
    0.01
    Negative Logits
    ema
    -0.73
    helm
    -0.72
     ENT
    -0.68
    mort
    -0.68
    catentry
    -0.67
    acker
    -0.67
    inventoryQuantity
    -0.63
    panic
    -0.63
    verson
    -0.63
    matter
    -0.63
    POSITIVE LOGITS
     Thumbnails
    0.83
    ãĥ«
    0.80
     Flickr
    0.79
    ãĥ´ãĤ¡
    0.77
    aina
    0.76
     Fighters
    0.73
    Tumblr
    0.72
    leon
    0.72
    heres
    0.70
    verning
    0.70
    Activations Density 0.290%

    No Known Activations