© 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
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 5521
    Prev
    Next
    INDEX
    Explanations

    instances of the word "just."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    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
    19
    +0.13
    0.6%
    430
    +0.11
    0.5%
    527
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    633
    +0.13
    0.05
    663
    +0.11
    0.04
    744
    +0.11
    0.05
    Negative Logits
     spoiler
    -0.67
    antage
    -0.67
     consideration
    -0.67
    emetery
    -0.62
    esting
    -0.61
     tolerant
    -0.61
     Celebrity
    -0.60
     complying
    -0.59
     spoilers
    -0.58
     Lauder
    -0.58
    POSITIVE LOGITS
    failed
    0.78
    ¬
    0.76
    Ĥİ
    0.74
    sych
    0.73
    ¸
    0.73
    tails
    0.72
    Ľ
    0.71
     shrugged
    0.71
    RT
    0.70
    Į
    0.70
    Activations Density 0.164%

    No Known Activations