AI Explained
Plain explanations of trending AI concepts, with live visualizations.
RELEX paper — Rank-1 RLVR weight-trajectory extrapolation — What does it mean?
RELEX exploits the empirical finding that RLVR fine-tuning weight trajectories are near rank-1 — fit a line through 15% of training steps and extrapolate; matches full RLVR quality from a fraction of the compute.
OScaR paper — Token Norm Imbalance — What does it mean?
Token Norm Imbalance — a few tokens carry outsized KV norms along the sequence axis. Channel rotation can't flatten them; OScaR can.
MSSP paper — Scale-stable parameterization beyond muP — What does it mean?
MSSP applies Dynamical Mean Field Theory to MoE training and derives a parameterization that — unlike muP — keeps the optimal learning rate stable as both model width and expert count scale.
Mix-Quant paper — NVFP4 prefill + BF16 decode — What does it mean?
Mix-Quant quantizes only the prefill phase to NVFP4 and keeps decode in BF16 — up to 3× prefill speedup with task performance largely preserved, because the two phases sit on opposite sides of the roofline.
PSD paper — Parallel speculative decoding for diffusion LLMs — What does it mean?
Parallel speculative decoding gets up to 5.5× tokens per forward pass on diffusion LLMs by attacking spatial and temporal axes at once.
OpenComputer paper — Verifier-grounded benchmark synthesis — What does it mean?
OpenComputer builds 1,000 computer-use tasks across 33 desktop apps by writing the executable verifier first, calibrating it, then synthesizing tasks that ground into the verifier endpoints — GPT-5.4 hits 68.3% while open-source agents collapse to as low as 5.7%.
MCP SEP-2106 — Full JSON Schema 2020-12 in tool I/O — What does it mean?
MCP SEP-2106 opens tool inputSchema and outputSchema to the full JSON Schema 2020-12 vocabulary — composition, conditionals, refs — and widens structuredContent from object-only to any value.
EnvFactory paper — Synthetic envs for tool-use agent training — What does it mean?
EnvFactory autonomously synthesizes 85 stateful tool environments — 5x fewer than EnvScaler / AWM — and uses topology-aware trajectory sampling to lift Qwen3 tool-use by up to 15 percentage points on BFCL v3.
Attention Once Is All You Need — Persistent KV cache across queries — What does it mean?
AOIAYN persists the KV cache across queries in a streaming session and advances it as data arrives — prefill leaves the critical path and per-query latency stays constant in context length.
ZEDA paper — Zero-output expert self-distillation — What does it mean?
ZEDA injects parameter-free zero-output experts into a finished MoE and uses two-stage self-distillation to teach the router to skip ~50% of expert FLOPs at marginal accuracy loss.
SGLang v0.5.12 — TokenSpeed MLA backend — What does it mean?
SGLang v0.5.12 ships TokenSpeed MLA — a Blackwell attention backend for Multi-head Latent Attention that caches one shared low-rank K/V latent instead of per-head K/V, with TMA bulk-store reporting up to ~12× speedup on the cache-write kernel.
MCP SEP-2468 — RFC 9207 iss parameter for OAuth mix-up defense — What does it mean?
SEP-2468 recommends MCP authorization servers include an iss parameter on auth responses, and requires clients to validate it string-equal against the recorded issuer — blocking OAuth mix-up across multi-IdP setups.