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. 6439
    Prev
    Next
    INDEX
    Explanations

    mentions of cats

    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
    336
    +0.20
    1.6%
    503
    +0.15
    1.2%
    160
    +0.15
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    167
    +0.20
    0.01
    651
    +0.15
    0.02
    36
    +0.15
    0.02
    Negative Logits
     confir
    -0.99
    nesday
    -0.93
    ynski
    -0.92
    ortium
    -0.90
    DragonMagazine
    -0.86
    avorite
    -0.85
    orsi
    -0.83
    xtap
    -0.83
    ignt
    -0.82
    icient
    -0.82
    POSITIVE LOGITS
    neys
    0.98
    abases
    0.84
    other
    0.78
    Minecraft
    0.78
    amins
    0.77
    alogue
    0.77
    Names
    0.75
    AMS
    0.74
    isha
    0.73
    bre
    0.73
    Activations Density 0.025%

    No Known Activations