DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models · HackerPapers discussions

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by thatsadude

Thu, 17 Sep 2026 06:20:04 GMT

https://arxiv.org/pdf/2402.03300 DeepseekMath was published many months before O1-preview.

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by nl

Wed, 21 Jan 2026 06:03:10 GMT

Not exactly sure what you are looking for here. That GRPO works? > Group Relative Policy Optimization (GRPO), a variant reinforcement learning (RL) algorithm of Proximal Policy Optimization (PPO) (Schulman et al., 2017). GRPO foregoes the critic model, instead estimating the baseline from group scores, significantly reducing training resources. By solely using a subset of English instruction tuning data, GRPO obtains a substantial improvement over the strong DeepSeekMath-Instruct, including both in-domain (GSM8K: 82.9% → 88.2%, MATH: 46.8% → 51.7%) and out-of-domain mathematical tasks (e.g.…

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by natrys

Thu, 27 Nov 2025 23:05:09 GMT

Well they do that too: https://huggingface.co/deepseek-ai/DeepSeek-Prover-V2-671B But I suppose the bigger goal remains improving their language model, and this was an experimentation born from that. These works are symbiotic; the original DeepSeekMath resulted in GRPO, which eventually formed the backbone of their R1 model: https://arxiv.org/abs/2402.03300

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by natrys

Wed, 14 May 2025 03:02:31 GMT

You could maybe make that accusation about V3 (to the extent that it's a bad thing and not fair use, specially considering amoral origin of OpenAI's models in first place), but don't think the claim makes sense for R1 since OpenAI's o1 did not expose its CoT traces even in API. They published about GRPO (key algorithm behind R1) a full year before[1] they scaled it for R1. Given the research they do in open, it's not far-fetched to think they had the talent and technical know-how to achieve R1 on their own. [1] https://arxiv.org/abs/2402.03300

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by layer8

Thu, 06 Mar 2025 22:16:24 GMT

GRPO = Group Relative Policy Optimization https://arxiv.org/abs/2402.03300

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by porridgeraisin

Thu, 27 Feb 2025 21:45:58 GMT

It is possible to understand the mechanism once you drop the anthropomorphisms. Each token output by an LLM involves one pass through the next-word predictor neural network. Each pass is a fixed amount of computation. Complexity theory hints to us that the problems which are "hard" for an LLM will need more compute than the ones which are "easy". Thus, the only mechanism through which an LLM can compute more and solve its "hard" problems is by outputting more tokens. You incentivise it to this end by human-grading its outputs ("RLHF") to prefer those where it spends time calculating before…

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by highfrequency

Tue, 28 Jan 2025 17:53:40 GMT

First graph tells the story - below a certain model size (500m params), reinforcement learning is close to useless. Above this (task-dependent) model size threshold, reinforcement learning basically works. I suspect this is what we saw play out with math/coding reasoning models - until recently, the base models were not good enough for ~random output search to hit on a correct path with any reasonable frequency. Below this threshold of base model intelligence, the only efficient way forward was to collect plain supervised data (either through human labeled math problem solutions [1] or…

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Tue, 28 Jan 2025 17:06:34 GMT

DeepSeekMath: Pushing Limits of Mathematical Reasoning in Open Language Models

1 point; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Sun, 26 Jan 2025 17:09:42 GMT

DeepSeekMath

6 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Wed, 21 Feb 2024 00:39:33 GMT

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open LLMs

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Tue, 20 Feb 2024 12:29:42 GMT

Pushing the limits of mathematical reasoning in open language models

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — mentioned by mdp

Tue, 06 Feb 2024 16:46:02 GMT

Related paper - https://arxiv.org/pdf/2402.03300.pdf

Paper and discussion history. A mention is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Tue, 06 Feb 2024 16:41:52 GMT

Pushing the Limits of Mathematical Reasoning in Open Language Models

2 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models — HN submission

Tue, 06 Feb 2024 04:45:22 GMT

Pushing the Limits of Mathematical Reasoning in Open Language Models

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.