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. 6436
    Prev
    Next
    INDEX
    Explanations

    variations of the word "rem" or related terms

    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
    503
    +0.17
    1.3%
    527
    +0.14
    1.0%
    706
    +0.14
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    503
    +0.17
    0.03
    444
    +0.14
    0.01
    389
    +0.14
    0.02
    Negative Logits
    rious
    -0.92
    tar
    -0.76
    warts
    -0.70
     Seymour
    -0.69
    tips
    -0.69
     Mour
    -0.69
    ument
    -0.67
    gerald
    -0.66
     Crown
    -0.66
    ãĥ¼ãĥĨãĤ£
    -0.65
    POSITIVE LOGITS
    isitions
    1.02
    abases
    0.95
    usalem
    0.94
    phas
    0.92
    hesda
    0.92
    osponsors
    0.91
    isition
    0.85
    endez
    0.85
    undai
    0.84
    apons
    0.83
    Activations Density 0.026%

    No Known Activations