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. 6443
    Prev
    Next
    INDEX
    Explanations

    words related to storytelling or narrative elements

    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
    0
    +0.18
    1.2%
    85
    +0.17
    1.1%
    57
    +0.15
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    26
    +0.18
    0.01
    206
    +0.17
    0.01
    228
    +0.15
    0.01
    Negative Logits
    archment
    -0.93
    çīĪ
    -0.81
    elsius
    -0.81
    ortium
    -0.79
    qqa
    -0.74
    renheit
    -0.74
    pmwiki
    -0.71
     mould
    -0.70
    âĵĺ
    -0.69
    perty
    -0.68
    POSITIVE LOGITS
    neys
    0.83
    bye
    0.76
    ossip
    0.75
     Smile
    0.73
    ctions
    0.71
    ote
    0.70
    ogi
    0.67
    miss
    0.66
    lehem
    0.66
    bors
    0.65
    Activations Density 0.008%

    No Known Activations