Unverified50% confidenceBenchmarkExact time
On six RULER needle-in-a-haystack tasks at a one-million-token setting, the paper reports that GDN's average accuracy improved from 1.7% to 70.3%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/15/2026
First Seen
Valid until: 1/13/2027
Sources
Related Entities
Related Claims
Unverified每步60%准确率意味着连续5步后整体成功率约为7.8%（0.6的5次方）67% similarUnverifiedSpeculative decoding technology makes the speed approximately 1.6 to 1.8 times faster, and in actual tests the large model accepted the small assistant's guesses approximately 79% of the time66% similarUnverifiedEvidence-driven pre-improvement validation improved precision from 58.6% to 72.3%, and reduced unsupported objectives per round from 1.93 to 0.50.66% similarUnverifiedIn the experiment, unit completion accuracy increased from 72.4% to 84.9%, a gain of 12.5 percentage points.65% similarUnverifiedAudit data published by the author princejain756 shows: 901 skills, 39,402 skill-description tokens, 15 MCP servers, 92,413 tokens burned per conversation turn, grade F.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1040228API
curl https://kongchang.com/api/v1/knowledge/claims/1040228MCP
get_claim(id=1040228)