[![Deep \(Learning\) Focus](https://substackcdn.com/image/fetch/$s_!87xa!,w_40,h_40,c_fill,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2Fab9b43fb-52d5-40da-995d-5b7cd3f91064_896x896.png)](https://cameronrwolfe.substack.com/)
# [Deep (Learning) Focus](https://cameronrwolfe.substack.com/)
SubscribeSign in
[Home](https://cameronrwolfe.substack.com/)
[Notes](https://cameronrwolfe.substack.com/notes)
[The Author](https://cameronrwolfe.me)
[Archive](https://cameronrwolfe.substack.com/archive)
[About](https://cameronrwolfe.substack.com/about)
## Page not found
LatestTopDiscussions
[Notes on NVIDIA Nemotron](https://cameronrwolfe.substack.com/p/nemotron)
[A broad coverage of the most practically useful series of open LLM tech reports...](https://cameronrwolfe.substack.com/p/nemotron)
Sep 28 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
47
9
![](../images/013.png)
### August 2026
[Reinforcement Learning for LLMs: The Complete Guide](https://cameronrwolfe.substack.com/p/llm-rl)
[Tracing the evolution of RL from first principles to the frontier of modern AI research...](https://cameronrwolfe.substack.com/p/llm-rl)
Aug 24 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
345
10
43
![](../images/031.png)
[Notes on Midtraining](https://cameronrwolfe.substack.com/p/midtraining-notes)
[Creating better specialized LLMs with midtraining and continual pretraining...](https://cameronrwolfe.substack.com/p/midtraining-notes)
Aug 10 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
73
8
![](../images/body-e3a473e183a6f8af92e4.png)
### July 2026
[Agentic World Models](https://cameronrwolfe.substack.com/p/agentic-world-models)
[Creating better language agents by teaching them to model their environment...](https://cameronrwolfe.substack.com/p/agentic-world-models)
Jul 20 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
128
9
19
![](../images/body-1012a3d330000e12e1e7.png)
### June 2026
[Agentic RL: Frameworks and Best Practices](https://cameronrwolfe.substack.com/p/agentic-rl)
[How LLMs are trained to handle long horizon tasks in complex environments...](https://cameronrwolfe.substack.com/p/agentic-rl)
Jun 22 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
194
14
20
![](../images/body-4c9994acaff529c793c6.png)
### May 2026
[Agent Evaluation: A Detailed Guide](https://cameronrwolfe.substack.com/p/agent-evals)
[Best practices and common patterns for effectively evaluating AI agents...](https://cameronrwolfe.substack.com/p/agent-evals)
May 18 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
338
26
53
![](../images/body-5cf22a261ce0c733924d.png)
### April 2026
[RL Scaling Laws for LLMs](https://cameronrwolfe.substack.com/p/rl-scaling-laws)
[How scaling laws have evolved from pretraining to reinforcement learning...](https://cameronrwolfe.substack.com/p/rl-scaling-laws)
Apr 20 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
144
3
17
![](../images/body-546fe271d6459d1e8944.png)
### March 2026
[The Anatomy of an LLM Benchmark](https://cameronrwolfe.substack.com/p/llm-bench)
[Common patterns used to create the most effective LLM evaluation datasets...](https://cameronrwolfe.substack.com/p/llm-bench)
Mar 30 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
118
4
17
![](../images/body-0c66677da0ae74552937.png)
[Applying Statistics to LLM Evaluations](https://cameronrwolfe.substack.com/p/stats-llm-evals)
[An overview of useful statistics for building and interpreting LLM evaluations...](https://cameronrwolfe.substack.com/p/stats-llm-evals)
Mar 9 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
148
9
15
![](../images/body-5488950eca7e4e671309.png)
### February 2026
[Rubric-Based Rewards for RL](https://cameronrwolfe.substack.com/p/rubric-rl)
[Extending the benefits of large-scale RL training to non-verifiable domains...](https://cameronrwolfe.substack.com/p/rubric-rl)
Feb 16
134
11
18
![](../images/body-41692c847a1946c758cc.png)
### January 2026
[Continual Learning with RL for LLMs](https://cameronrwolfe.substack.com/p/rl-continual-learning)
[Exploring the impressive continual learning capabilities of RL training...](https://cameronrwolfe.substack.com/p/rl-continual-learning)
Jan 26 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
160
15
20
![](../images/body-704df4e08e7d36e517be.png)
[GRPO++: Tricks for Making RL Actually Work](https://cameronrwolfe.substack.com/p/grpo-tricks)
[How to go from the vanilla GRPO algorithm to functional RL training at scale...](https://cameronrwolfe.substack.com/p/grpo-tricks)
Jan 5 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
167
10
18
![](../images/body-03fa8cffb92ee3176432.png)
### December 2025
[Olmo 3 and the Open LLM Renaissance](https://cameronrwolfe.substack.com/p/olmo-3)
[Fully-open artifacts with the potential to make LLM research a reality for anyone...](https://cameronrwolfe.substack.com/p/olmo-3)
Dec 15, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
85
7
14
![](../images/body-b5870b795892abcd6ba2.png)
### November 2025
[Group Relative Policy Optimization (GRPO)](https://cameronrwolfe.substack.com/p/grpo)
[How the algorithm that teaches LLMs to reason actually works...](https://cameronrwolfe.substack.com/p/grpo)
Nov 24, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
136
11
14
![](../images/body-3b7875e64c0be40c5c27.png)
### October 2025
[PPO for LLMs: A Guide for Normal People](https://cameronrwolfe.substack.com/p/ppo-llm)
[Understanding the complex RL algorithm that gave us modern LLMs…](https://cameronrwolfe.substack.com/p/ppo-llm)
Oct 27, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
185
13
14
![](../images/body-d376c3f6174436bf5e60.png)
### September 2025
[REINFORCE: Easy Online RL for LLMs](https://cameronrwolfe.substack.com/p/reinforce)
[How to get the benefits of online RL without the complexity of PPO...](https://cameronrwolfe.substack.com/p/reinforce)
Sep 29, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
106
11
6
![](../images/body-18445455c8f92b714e39.png)
[Online versus Offline RL for LLMs](https://cameronrwolfe.substack.com/p/online-rl)
[A deep dive into the online-offline performance gap in LLM alignment...](https://cameronrwolfe.substack.com/p/online-rl)
Sep 8, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
93
4
10
![](../images/body-e5907245629c928f8db4.png)
### August 2025
[GPT-oss from the Ground Up](https://cameronrwolfe.substack.com/p/gpt-oss)
[Everything you should know about OpenAI's new open-weight language models...](https://cameronrwolfe.substack.com/p/gpt-oss)
Aug 18, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
105
14
15
![](../images/body-631803b0aad1f4333c95.png)
### July 2025
[Direct Preference Optimization (DPO)](https://cameronrwolfe.substack.com/p/direct-preference-optimization)
[How to align LLMs with limited hardware and minimal complexity...](https://cameronrwolfe.substack.com/p/direct-preference-optimization)
Jul 28, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
142
20
11
![](../images/body-7150bc39909d5b48847f.png)
### June 2025
[Reward Models](https://cameronrwolfe.substack.com/p/reward-models)
[Modeling human preferences for LLMs in the age of reasoning models...](https://cameronrwolfe.substack.com/p/reward-models)
Jun 30, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
145
13
20
![](../images/body-cf341d25c37d3cedaf96.png)
[AI Agents from First Principles](https://cameronrwolfe.substack.com/p/ai-agents)
[Understanding AI agents by building upon the most basic concepts of LLMs...](https://cameronrwolfe.substack.com/p/ai-agents)
Jun 9, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
381
26
46
![](../images/body-ced1fdf7e22892eecc47.png)
### May 2025
[A Guide for Debugging LLM Training Data](https://cameronrwolfe.substack.com/p/llm-debugging)
[Data-centric techniques and tools that anyone should use when training an LLM...](https://cameronrwolfe.substack.com/p/llm-debugging)
May 19, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
89
3
13
![](../images/body-25c4f69725ccd3ac19ae.png)
### April 2025
[Llama 4: The Challenges of Creating a Frontier-Level LLM](https://cameronrwolfe.substack.com/p/llama-4)
[The full story behind Llama 4 and Meta's huge pivot in research strategy...](https://cameronrwolfe.substack.com/p/llama-4)
Apr 28, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
83
2
11
![](../images/body-7f695f890ad621e1b37d.png)
### March 2025
[Vision Large Language Models (vLLMs)](https://cameronrwolfe.substack.com/p/vision-llms)
[Teaching LLMs to understand images and videos in addition to text...](https://cameronrwolfe.substack.com/p/vision-llms)
Mar 31, 2025 • [Cameron R. Wolfe, Ph.D.](https://substack.com/@cwolferesearch)
156
15
14
![](../images/body-3cde6af56e6afa149880.png)
© 2026 Cameron R. Wolfe · [Privacy](https://substack.com/privacy) ∙ [Terms](https://substack.com/tos) ∙ [Collection notice](https://substack.com/ccpa#personal-data-collected)
[ Start your Substack](https://substack.com/signup?utm_source=substack&utm_medium=web&utm_content=footer)[Get the app](https://substack.com/app/app-store-redirect?utm_campaign=app-marketing&utm_content=web-footer-button)
[Substack](https://substack.com) is the home for great culture