🛰️ Daily AI Frontier
‹ back to 2026-09-10

TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards

Research LLM Agents

Ranking

Overall 81
Content 100
Popularity 37

Observed public metrics from 1 member.

Representative image for TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards

Merged summary

TL;DR - TRACE trains tool-using reasoning agents to diagnose simulated digital-advertising anomalies using objective rewards derived from hidden interventions. Combining supervised fine-tuning with reinforcement learning improved accuracy beyond prompted frontier models while reducing tool use.

  • The simulator injects one of 12 root causes, producing noisy and confounded observations while retaining an oracle label for verifiable rewards.
  • Agents use Python and SQL to identify both the root cause and, when relevant, the affected segment.
  • On 235 held-out episodes, Qwen3.5-35B-A3B improved from 0.159 FullAttr@1 to 0.637 after supervised fine-tuning and 0.757 after reinforcement learning.
  • The RL-trained model outperformed the strongest prompted baseline, Claude Opus 5 at 0.686, suggesting scalable verification can matter more than model size alone.

Sources (1)

TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards

arXiv cs.AI Rui Sun, Zhan Shi, Bing He 2026-09-09 arXiv:2609.10315
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-26 14:16:26.608645 UTC

TL;DR - TRACE trains tool-using reasoning agents to diagnose simulated digital-advertising anomalies using objective rewards derived from hidden interventions. Combining supervised fine-tuning with reinforcement learning improved accuracy beyond prompted frontier models while reducing tool use.

  • The simulator injects one of 12 root causes, producing noisy and confounded observations while retaining an oracle label for verifiable rewards.
  • Agents use Python and SQL to identify both the root cause and, when relevant, the affected segment.
  • On 235 held-out episodes, Qwen3.5-35B-A3B improved from 0.159 FullAttr@1 to 0.637 after supervised fine-tuning and 0.757 after reinforcement learning.
  • The RL-trained model outperformed the strongest prompted baseline, Claude Opus 5 at 0.686, suggesting scalable verification can matter more than model size alone.
item →