10 items with this tag.
frontiers
Bleeding-edge synthesis pages for the most actionable recent seams in compact-model research for Parameter Golf.
frontiers
Frontier synthesis on the newest shift from handcrafted post-training formats toward training rules and learned representations that directly target compressible model weights.
frontiers
Frontier synthesis on why the best next compression wins may come from explicit byte-return accounting rather than better nominal low-bit methods alone.
challenge history
How Parameter Golf is already shifting from a tiny-model challenge to an artifact-aware systems-and-modeling problem, even with only a thin public run record.
frontiers
Frontier synthesis on why the next gains may come from allocating scarce high-fidelity bytes intelligently rather than chasing a single global quantization regime.
frontiers
Frontier synthesis on why recursive models may only become compelling once normalization and tiny phase-specific adaptation are treated as part of the compression interface.
frontiers
Frontier synthesis on why repeated structure, clustered values, and regular exception patterns may matter more than nominal precision once the final artifact and metadata are counted.
frontiers
Frontier synthesis on whether bounded evaluation-time compute can reconstruct some of the capacity that a hard artifact cap prevents us from storing directly.
frontiers
Frontier synthesis on why tokenizer research for compact models should be treated as a joint vocabulary, logits, and artifact-budget problem rather than a token-count problem.
hypotheses
Synthesis hypothesis that the strongest compact artifacts will combine shared depth, activation discipline, selective precision, and cheap specialization rather than relying on one trick alone.