mlsanityData to 14 Oct 2026

Soft Teacher Forcing

arXiv 2610.17189 ↗First posted 12 Oct 2026cs.CL cs.LGcited by 0

On the map: LLM reasoning → Test-time compute and efficient thinking also fits: Prompting and chain-of-thought reasoning

Abstract (arXiv)

Soft tokens have so far been used to replace discrete chain-of-thought in latent reasoning methods. These approaches typically require multi-stage training and, in some cases, auxiliary modules, yet they still underperform supervised fine-tuning (SFT) on <input, chain-of-thought, output> triples. In this paper, we ask whether soft tokens can be used to improve SFT itself, instead of as an avenue for latent reasoning. We introduce Soft Teacher Forcing (Soft-TF), a simple training algorithm that combines teacher-forced embeddings with soft tokens through random convex combinations. The soft tokens are computed in a single parallel forward pass over the ground-truth sequence. Soft-TF adds no parameters, no external model and no curriculum, and leaves inference unchanged. Experiments show that Soft-TF improves over SFT on GSM8K, MATH, MathQA and AQuA, by +2.1 to +4.1 points on GSM8K across seven Gemma, Llama and Mistral models.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 34 of its 41 references that are in the map. First posted on arXiv 12 Oct 2026.

Builds on

34 of 41 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoning5 Feb

Other references · 33

  1. 2017Attention Is All You Needcited by 60,361
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  4. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  5. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  6. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  7. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  8. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  9. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  10. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  11. 2023Let's Verify Step by Stepcited by 3,789
  12. 2022Scaling Instruction-Finetuned Language Modelscited by 3,024
21 more
  1. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
  2. 2025Gemma 3 Technical Reportcited by 1,824
  3. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  4. 2018Adafactor: Adaptive Learning Rates with Sublinear Memory Costcited by 905
  5. 2015Scheduled Sampling for Sequence Prediction with Recurrent Neural Networkscited by 803
  6. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  7. 2019MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalismscited by 662
  8. 2017Program Induction by Rationale Generation : Learning to Solve and Explain Algebraic Word Problemscited by 650
  9. 2021Domain Generalization with MixStylecited by 566
  10. 2023MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuningcited by 487
  11. 2017Regularizing Neural Networks by Penalizing Confident Output Distributionscited by 464
  12. 2023Orca: Progressive Learning from Complex Explanation Traces of GPT-4cited by 379
  13. 2024From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Stepcited by 235
  14. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 225
  15. 2025Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Spacecited by 117
  16. 2025SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMscited by 116
  17. 2023NEFTune: Noisy Embeddings Improve Instruction Finetuningcited by 100
  18. 2025SIM-CoT: Supervised Implicit Chain-of-Thoughtcited by 70
  19. 2019Scheduled Sampling for Transformerscited by 48
  20. 2025LLMs are Single-threaded Reasoners: Demystifying the Working Mechanism of Soft Thinkingcited by 38
  21. 2025Soft Tokens, Hard Truthscited by 35

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Improving Latent Reasoning in LLMs via Soft Concept Mixing
  2. 2026SoFT: Soft Targets for Generalizable LLM Fine-Tuning
  3. 2026Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning
  4. 2025SoftCoT++: Test-Time Scaling with Soft Chain-of-Thought Reasoning
  5. 2026From Noise to Diversity: Random Embedding Injection in LLM Reasoning
  6. 2026SeLaR: Selective Latent Reasoning in Large Language Models
  7. 2026One Student, Many Teachers: Multi-Task On-Policy Distillation via Soft-Prompt Privileged Context
  8. 2026RuPLaR : Efficient Latent Compression of LLM Reasoning Chains with Rule-Based Priors From Multi-Step to One-Step