Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? · HackerPapers discussions

Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? — mentioned by thesz

Wed, 29 Jul 2026 19:57:12 GMT

> The tokens it "predicts" aren't sampled from any naturally occurring distribution; the model's output is the result of an optimisation process that rewarded behaviour that was useful, and that's fundamentally different. https://arxiv.org/abs/2504.13837 "Surprisingly, we find that the current training setup does not elicit fundamentally new reasoning patterns. While RLVR-trained models outperform their base models at small k (e.g., k = 1), the base models achieve a higher pass@k score when k is large. Coverage and perplexity analyses show that the observed reasoning abilities originate from…

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

Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? — mentioned by hodgehog11

Sat, 28 Mar 2026 22:58:02 GMT

This argument, that LLMs can develop new crazy strategies using RLVR on math problems (like what happened with Chess), turns out to be false without a serious paradigm shift. Essentially, the search space is far too large, and the model will need help to explore better, probably with human feedback. https://arxiv.org/abs/2504.13837

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

Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? — HN submission

Tue, 24 Jun 2025 12:37:05 GMT

Does Reinforcement Learning Incentivize Reasoning Capacity in LLMs?

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

Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? — mentioned by anakin87

Sat, 10 May 2025 09:26:34 GMT

I experimented with GRPO lately, since I am fascinated by models learning from prompts and rewards - no example answers needed like in Supervised Fine-Tuning. After the DeepSeek boom, everyone is trying GRPO with GSM8K or the Countdown Game, but I wanted a different challenge. So I opted for teaching a model to create a schedule from a list of events and priorities. Choosing an original problem forced me to think about the problem setting, generate data, choose the base model, design reward functions, and run multiple rounds of training, hoping that my model would learn something. A fun and…

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

Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? — HN submission

Thu, 24 Apr 2025 12:57:34 GMT

Does RL Incentivize Reasoning Capacity in LLMs Beyond the Base Model?

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