Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.09621 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911440857726976 |
|---|---|
| author | Lyngkhoi, R E Zera Marveen Chawla, Chirag Seth, Pratinav Avaiya, Utsav Bhattacharjee, Soham Khandoga, Mykola Yuan, Rui Sankarapu, Vinay Kumar |
| author_facet | Lyngkhoi, R E Zera Marveen Chawla, Chirag Seth, Pratinav Avaiya, Utsav Bhattacharjee, Soham Khandoga, Mykola Yuan, Rui Sankarapu, Vinay Kumar |
| contents | Post-training alignment is central to deploying large language models (LLMs), yet practical workflows remain split across backend-specific tools and ad-hoc glue code, making experiments hard to reproduce. We identify backend interference, reward fragmentation, and irreproducible pipelines as key obstacles in alignment research. We introduce AlignTune, a modular toolkit exposing a unified interface for supervised fine-tuning (SFT) and RLHF-style optimization with interchangeable TRL and Unsloth backends. AlignTune standardizes configuration, provides an extensible reward layer (rule-based and learned), and integrates evaluation over standard benchmarks and custom tasks. By isolating backend-specific logic behind a single factory boundary, AlignTune enables controlled comparisons and reproducible alignment experiments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_09621 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models Lyngkhoi, R E Zera Marveen Chawla, Chirag Seth, Pratinav Avaiya, Utsav Bhattacharjee, Soham Khandoga, Mykola Yuan, Rui Sankarapu, Vinay Kumar Computation and Language Machine Learning Post-training alignment is central to deploying large language models (LLMs), yet practical workflows remain split across backend-specific tools and ad-hoc glue code, making experiments hard to reproduce. We identify backend interference, reward fragmentation, and irreproducible pipelines as key obstacles in alignment research. We introduce AlignTune, a modular toolkit exposing a unified interface for supervised fine-tuning (SFT) and RLHF-style optimization with interchangeable TRL and Unsloth backends. AlignTune standardizes configuration, provides an extensible reward layer (rule-based and learned), and integrates evaluation over standard benchmarks and custom tasks. By isolating backend-specific logic behind a single factory boundary, AlignTune enables controlled comparisons and reproducible alignment experiments. |
| title | AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2602.09621 |