© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Attention Out - 32k
    5. 3-ATT_32K-OAI
    6. 3545
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate relationships, changes, or conditions involving entities and their actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-att_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-attn-out-SAEs/v5_32k_layer_3
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.3.hook_attn_out
    Hook Layer
    3
    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
    672
    +0.14
    0.7%
    738
    +0.12
    0.6%
    91
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    672
    +0.14
    0.08
    736
    +0.12
    0.08
    450
    +0.11
    0.07
    Negative Logits
    ategory
    -0.86
    ascript
    -0.74
    etz
    -0.73
    berra
    -0.72
    achu
    -0.70
    atan
    -0.66
    egal
    -0.65
    lace
    -0.65
    erala
    -0.65
    rica
    -0.64
    POSITIVE LOGITS
     Claims
    0.66
     Pry
    0.65
     couldn
    0.64
     heir
    0.63
     Monkey
    0.62
     NEC
    0.61
     Alert
    0.60
    heimer
    0.59
    ')
    0.59
     Scrib
    0.57
    Activations Density 1.655%

    No Known Activations