TIER creates trajectory-invariant rewards from tool schemas and execution results for multi-step LLM tool use, reaching over 90 percent accuracy on DepthBench up to six steps where reference-trajectory methods fail after four.
Wait for the environment to execute it and provide feedback before proceeding with the next step
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool Composition
TIER creates trajectory-invariant rewards from tool schemas and execution results for multi-step LLM tool use, reaching over 90 percent accuracy on DepthBench up to six steps where reference-trajectory methods fail after four.