Unverified50% confidenceBenchmarkExact time
In comparison with SFT, JET achieves 18% to 28% fewer tool calls than SFT on expert trajectories while retaining most of SFT's score improvement.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/18/2026
First Seen
Valid until: 2/16/2027
Sources
Related Entities
Related Claims
UnverifiedThe paper evaluates Ambient SFT on domain adaptation, instruction fine-tuning and reasoning tasks, and the results show that the method outperforms standard SFT in all these scenarios, with an average relative improvement of 21%.71% similarUnverifiedOn self-generated trajectories, JET reduces the number of tool calls by 11% to 33% while largely maintaining task scores.67% similarUnverifiedIn the SST-2 dataset DistilBERT-to-DistilBERT self-distillation experiment, the reliability gating method achieved 88.800%, RAPID achieved 88.685%, representing an improvement of approximately 1.4 percentage points over the baseline of 87.271%.64% similarUnverifiedDuring training, the held-out FID dropped from 33.7 to 27.0, FD-DINOv2 dropped from 570 to 218, and detector-based object accuracy improved from 65% to 90%.61% similarUnverifiedControlled experiments show that after providing models with complete requirements, the F2P pass rates of Seed Evolving and GPT-6 Astra increased by about 16 percentage points, but the full task success rate did not change at all.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1182987API
curl https://kongchang.com/api/v1/knowledge/claims/1182987MCP
get_claim(id=1182987)