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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 5564
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "cons" in various contexts, likely indicating references to "conservative" sentiments or attitudes

    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
    463
    +0.10
    0.5%
    472
    +0.10
    0.4%
    503
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    154
    +0.10
    0.01
    536
    +0.10
    0.01
    555
    +0.10
    0.01
    Negative Logits
    iencies
    -0.72
    ateurs
    -0.69
    usters
    -0.68
    emetery
    -0.67
    ussen
    -0.63
    pants
    -0.63
    orem
    -0.63
    uses
    -0.62
    ewski
    -0.62
    aird
    -0.61
    POSITIVE LOGITS
    ãĥ«
    1.02
    ãĥĭ
    0.98
    ãĥ³ãĤ¸
    0.91
    ãĥİ
    0.91
    ãĥ©ãĥ³
    0.85
    çĶ°
    0.84
    ãĥķãĤ¡
    0.82
    ãĥ´
    0.81
    ãĤ±
    0.78
    ãĤ¸
    0.78
    Activations Density 0.007%

    No Known Activations