← All tracks

Attention Variants

After Attention 101 — the real-world variants that actually run in modern LLMs.

0 / 14 solved Continue →
  1. 1. ○ Cross Attention
  2. 2. ○ Causal Attention Mask
  3. 3. ○ Grouped-Query Attention
  4. 4. ○ Sliding Window Attention
  5. 5. ○ Efficient Attention with Masking
  6. 6. ○ KV Cache for Autoregressive Decoding
  7. 7. ○ Flash Attention Score Computation
  8. 8. ○ Causal Self-Attention Block
  9. 9. ○ Cross-Attention Block
  10. 10. ○ LLaMA-Style Transformer Block
  11. 11. ○ Encoder-Decoder Transformer Forward Pass
  12. 12. ○ Train Encoder-Decoder Seq2Seq Step
  13. 13. ○ Encoder-Decoder Greedy Decode
  14. 14. ○ Encoder-Decoder Beam Search