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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 13571
    Prev
    Next
    INDEX
    Explanations

    occurrences of the letter 'M' in various contexts

    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
    167
    +0.31
    2.3%
    57
    +0.19
    1.4%
    718
    +0.17
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    57
    +0.31
    0.06
    659
    +0.19
    0.03
    746
    +0.17
    0.04
    Negative Logits
    DragonMagazine
    -1.06
    âķIJ
    -1.01
     Converted
    -0.84
    ;;;;;;;;;;;;
    -0.82
    mobi
    -0.81
    =-=-
    -0.81
     Flavoring
    -0.80
     Tradable
    -0.78
     Freddie
    -0.77
    PsyNetMessage
    -0.76
    POSITIVE LOGITS
     corrid
    0.97
     resil
    0.96
    ntil
    0.93
     contrace
    0.92
     metic
    0.91
    izoph
    0.90
     nodd
    0.86
    yond
    0.85
     submar
    0.83
    esty
    0.83
    Activations Density 0.138%

    No Known Activations