Trained on the retail domain of tau2-bench alone rather than on airline+retail jointly, completing a four-way sweep (plain/memory x airline/retail). On retail the memory-augmented variant is the stronger of the two at this budget (avg@3 0.558 vs 0.492, pass^3 0.475 vs 0.350), and that ordering also held under joint training. Joint training carried to 5 epochs reached avg@3 0.592 (memory) and 0.558 (plain), so epoch count mattered more than domain separation here. Numbers use a 60-hop episode cap rather than tau2's 200-hop default, so they are not comparable to the public leaderboard.
Independent publisher
MinseokKang
MANGSEOK123
Models in Library2
Datasets in Library0
Models on Hugging Face4
Followers—
Compare the published evidence
This profile connects the 2 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Reinforcement learning.
Reported licenses: apache-2.0. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Models
Qwen3-4B-Instruct-2507 consolidated with OEL on the tau2-bench telecom domain, using 120 externally authored task–memory pairs from the revised finalpairs2 set, stratified down from 125 so the batches divide evenly. The student replays each task without memory; the teacher is the same weights with that task's memory in the system prompt. Only the prompt differs, and no reward is used. Five pairs were dropped from the 125 by stratified sampling (seed 42), holding the level ratios to within 1pp, so that no batch is short — earlier runs showed a tiny final batch wrecks the saved checkpoint, and only the last step is saved. Each step reads a different batch, so the loss column tracks batch…