MIT

Trends

• Efficiency methods cut the compute and memory models need

Items

1. Oct 9

   MIT's fault-tolerant foundation models

   LLMs trained with the same hardware faults they will meet at inference grow more error-resilient as they scale, following a modified neural scaling law that points to running inference on low-energy, unreliable hardware for large energy savings.

2. Apr 6

   TriAttention compresses the KV cache 10.7x

   A training-free method from MIT, NVIDIA, and ZJU scores KV importance from the stable trigonometric structure of pre-RoPE queries and keys, matching full attention on AIME25 at 2.5x throughput or 10.7x less KV memory and fitting a 32B reasoning model on a 24GB RTX 4090.