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. 13554
    Prev
    Next
    INDEX
    Explanations

    mentions of reading and the act of reading

    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
    6
    +0.21
    1.5%
    642
    +0.18
    1.3%
    26
    +0.16
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    26
    +0.21
    0.06
    568
    +0.18
    0.03
    615
    +0.16
    0.03
    Negative Logits
     millenn
    -1.12
    ccording
    -1.03
    ahon
    -1.01
    ategory
    -1.01
     veter
    -0.99
    regor
    -0.98
    avorite
    -0.94
    ascus
    -0.93
    ¥ŀ
    -0.92
    undai
    -0.91
    POSITIVE LOGITS
     IPM
    0.81
    \\\\\\\\
    0.80
     Rainbow
    0.76
    ç·
    0.75
     aloud
    0.72
     Illum
    0.71
    \\\\\\\\\\\\\\\\
    0.69
     Sakuya
    0.68
    recomm
    0.68
     annotations
    0.67
    Activations Density 0.084%

    No Known Activations