Playground

Move the dials — chunk size, top-k, reranking, hybrid search — and watch retrieval quality and latency shift.

Shard a model across GPUs with tensor, pipeline, and data parallelism plus ZeRO — compare illustrative state-memory estimates.

Fit one model to many random samples and watch the fits scatter (variance) and drift from truth (bias) as complexity, data, and noise change.

Play four arms with greedy, ε-greedy, and UCB on the same reward draws — and watch which strategies keep cumulative regret low.

Score a group of responses and watch GRPO turn rewards into advantages using the group's own mean as the baseline — no critic, and it collapses when rewards agree.

Drag the importance ratio and flip the advantage's sign to see how PPO clipping saturates part of the surrogate objective; it does not enforce a hard ratio or KL bound.

Pick a query token and inspect softmax(QKᵀ/√d_k). Adjust temperature to compare concentrated and spread-out attention, or mask future tokens for autoregression.

Compare modeled attention links, score storage and score-matmul work as a sequence grows. See how fixed windows and sparse patterns change the scaling.

Hold query and key content fixed, then rotate them with position. See why their RoPE dot product depends on the relative position gap.

Estimate weight and KV-cache storage for a model preset. Compare hypothetical MHA, GQA, MQA and MLA layouts, context lengths and cache precisions.

Try synthetic routing patterns and inspect expert load, capacity and dropped routes. Compare top-k choices and an illustrative balancing adjustment.

Walk one Model Context Protocol session — connect, negotiate, discover tools, call one, feed the result back — and see how a poisoned tool description or an over-broad permission changes the outcome.