Metraipolis › archive › dc_601b5abd6839

Inverse Reasoning Models for Efficient Post-Training and Verification from Answer-Only Data — arXiv:2610.18878

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.18878
License (as archived)
CC0-1.0
Archived
2026-10-15T14:14:31.220Z
SHA-256 of the archived text
ecf5f9212833d2a5e654cdee08bf88a2c8cb43604f8d8bd6abe52a1228e1c439
Stable URL of this page
https://metraipolis.com/documento/dc_601b5abd6839

How to cite

Inverse Reasoning Models for Efficient Post-Training and Verification from Answer-Only Data — arXiv:2610.18878 — arXiv (en), https://arxiv.org/abs/2610.18878 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_601b5abd6839 (Metraipolis archive dc_601b5abd6839, sha256 ecf5f9212833d2a5e654cdee08bf88a2c8cb43604f8d8bd6abe52a1228e1c439)
BibTeX
@misc{dc601b5abd6839,
  title = {Inverse Reasoning Models for Efficient Post-Training and Verification from Answer-Only Data — arXiv:2610.18878},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.18878},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_601b5abd6839; sha256 ecf5f9212833d2a5e654cdee08bf88a2c8cb43604f8d8bd6abe52a1228e1c439},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.18878. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Post-training reasoning LLMs often requires making a tradeoff between costly supervision and compute inefficiency: supervised finetuning typically requires reasoning traces, which is data inefficient since in many applications one only has prompt-answer tuples and lacks the groundtruth chain-of-thought (CoT) sequence. On the other hand, reinforcement learning (with verifiable rewards, for example) can operate without CoT data but it is much more compute-inefficient since it requires multiple autoregressive completions per optimization step, none of which are guaranteed to give a non-zero policy gradient. In this work we explore a middleground, a self-distillation technique that uses answer-only data, treating CoT as latent variables. This relies on training a partner ``inverse reasoning LLM" that generates the original's reasoning traces conditioned on the answer. The inverse model is trained once and reused across incoming batches of answer-only data, with all generation happening offline and up front, making the approach particularly suited to continual post-training. Using Qwen3-0.6B on the Countdown and GSM8K tasks, we show that after GRPO training has saturated, a single offline distillation round from the inverse model roughly doubles pass@1 on hard prompts and, on Countdown, outperforms rejection-sampling self-distillation on every hard-tail metric at every mix ratio (with smaller gains on GSM8K); rejection sampling by construction has no signal on problems the model never solves on its own. Continued GRPO on the fresh data retains an edge at hard pass@1 but trails distillation at pass@8 and pass@32 while regenerating rollouts at every step. As an added bonus we see that the inverse model has a second application as an accurate verifier of correctness. When used for best-of-$n$ selection at test time, it raises selection accuracy on Countdown from 0.753 (majority voting) to 0.969.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.18878 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T13:53:35.487Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).