Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning
Paper
β’
2502.18080
β’
Published
β’
2
Junlin Zhou
jlzhou
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
1 day ago
Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning
upvoted
an
article
2 days ago
Open R1: Update #3
liked
a model
2 days ago
open-r1/OlympicCoder-32B
Organizations
jlzhou's activity
upvoted
a
paper
1 day ago
upvoted
an
article
2 days ago
Article
Open R1: Update #3
By
and 9 others
β’
β’
197
liked
a
model
2 days ago
liked
a
dataset
2 days ago
upvoted
an
article
4 days ago
Article
From Files to Chunks: Improving Hugging Face Storage Efficiency
β’
52
reacted to
Kseniase's
post with π
11 days ago
Post
6071
9 types of "Chain-of-..." approaches:
Chain-of-Thought (CoT) prompting enhances reasoning in AI models by breaking down complex problems into step-by-step logical sequences. It continues proving its effectiveness, especially in top-performing reasoning models. However, there are other similar methods, that expand CoT and can be used for different purposes. Here are 9 of them:
1. Chain-of-Action-Thought (COAT) -> Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search (2502.02508)
Helps model decide when to keep thinking, double-check their work, or try a different approach, using special guiding tokens.
2. Chain of Draft (CoD) -> Chain of Draft: Thinking Faster by Writing Less (2502.18600)
It helps model generate short but meaningful reasoning steps, cutting costs and making processing faster
3. Chain-of-Agents -> Chain of Agents: Large Language Models Collaborating on Long-Context Tasks (2406.02818)
Uses multi-agent collaboration: Worker agents process text parts in a structured chain, and manager agent summarizes the results
4. Chain-of-RAG ->https://huggingface.co/papers/2501.14342
Creates retrieval chains, instead of retrieving all info at once. It can dynamically adjust its search process and its parameters like step number
5. Chain-of-Shot Prompting (CoS) -> CoS: Chain-of-Shot Prompting for Long Video Understanding (2502.06428)
Helps models pick frames crucial for understanding a video, using a binary video summary and video co-reasoning module.
6. Chain of Hindsight (CoH) -> Chain of Hindsight Aligns Language Models with Feedback (2302.02676)
Converts all feedback into sequences to fine-tune the model and refine outputs
7. Chain-of-Note (CoN) -> Chain-of-Note: Enhancing Robustness in Retrieval-Augmented Language Models (2311.09210)
Generates sequential reading notes for each retrieved document to assess relevance before integrating info into the final answer
8. Chain of Diagnosis (CoD) -> CoD, Towards an Interpretable Medical Agent using Chain of Diagnosis (2407.13301)
Transforms the diagnostic process into a diagnostic chain
9. Chain(s)-of-Knowledge -> https://www.turingpost.com/p/cok
Enhance LLMs by dynamically pulling in external knowledge to improve accuracy and reduce errors
Chain-of-Thought (CoT) prompting enhances reasoning in AI models by breaking down complex problems into step-by-step logical sequences. It continues proving its effectiveness, especially in top-performing reasoning models. However, there are other similar methods, that expand CoT and can be used for different purposes. Here are 9 of them:
1. Chain-of-Action-Thought (COAT) -> Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search (2502.02508)
Helps model decide when to keep thinking, double-check their work, or try a different approach, using special guiding tokens.
2. Chain of Draft (CoD) -> Chain of Draft: Thinking Faster by Writing Less (2502.18600)
It helps model generate short but meaningful reasoning steps, cutting costs and making processing faster
3. Chain-of-Agents -> Chain of Agents: Large Language Models Collaborating on Long-Context Tasks (2406.02818)
Uses multi-agent collaboration: Worker agents process text parts in a structured chain, and manager agent summarizes the results
4. Chain-of-RAG ->https://huggingface.co/papers/2501.14342
Creates retrieval chains, instead of retrieving all info at once. It can dynamically adjust its search process and its parameters like step number
5. Chain-of-Shot Prompting (CoS) -> CoS: Chain-of-Shot Prompting for Long Video Understanding (2502.06428)
Helps models pick frames crucial for understanding a video, using a binary video summary and video co-reasoning module.
6. Chain of Hindsight (CoH) -> Chain of Hindsight Aligns Language Models with Feedback (2302.02676)
Converts all feedback into sequences to fine-tune the model and refine outputs
7. Chain-of-Note (CoN) -> Chain-of-Note: Enhancing Robustness in Retrieval-Augmented Language Models (2311.09210)
Generates sequential reading notes for each retrieved document to assess relevance before integrating info into the final answer
8. Chain of Diagnosis (CoD) -> CoD, Towards an Interpretable Medical Agent using Chain of Diagnosis (2407.13301)
Transforms the diagnostic process into a diagnostic chain
9. Chain(s)-of-Knowledge -> https://www.turingpost.com/p/cok
Enhance LLMs by dynamically pulling in external knowledge to improve accuracy and reduce errors
liked
a
model
17 days ago
liked
a
dataset
21 days ago
commented on
Distributed SFT with trl and DeepSpeed Part 2: Scaling Locally
23 days ago
I'm glad you found it helpful!
Yes, this is planned. I was originally planning to write an article about training with the training operator, but now I'm wondering if I should skip that and focus on training with the new trainer instead.
PS: Kubeflow is migrating their training component from v1 (Kubeflow Training Operator) to v2 (Kubeflow Trainer).
upvoted
a
paper
25 days ago
liked
a
dataset
25 days ago
upvoted
a
paper
28 days ago
updated
a
model
29 days ago
updated
a
collection
29 days ago
commented
a
paper
about 1 month ago
reacted to
schuler's
post with π
about 1 month ago
Post
7226
π’ New Research Alert: Making Language Models Smaller & Smarter!
Thrilled to share the latest technical report demonstrating how to reduce language model parameters by 77% while maintaining performance.
The secret? Grouped pointwise convolutions. Yes. We brought a method from computer vision to the transformers arena.
π Key Findings:
β’ 77% parameter reduction.
β’ Maintained model capabilities.
β’ Improved generalization.
Paper: https://www.researchgate.net/publication/388835829_SAVING_77_OF_THE_PARAMETERS_IN_LARGE_LANGUAGE_MODELS_TECHNICAL_REPORT
Code: https://github.com/joaopauloschuler/less-parameters-llm
Thrilled to share the latest technical report demonstrating how to reduce language model parameters by 77% while maintaining performance.
The secret? Grouped pointwise convolutions. Yes. We brought a method from computer vision to the transformers arena.
π Key Findings:
β’ 77% parameter reduction.
β’ Maintained model capabilities.
β’ Improved generalization.
Paper: https://www.researchgate.net/publication/388835829_SAVING_77_OF_THE_PARAMETERS_IN_LARGE_LANGUAGE_MODELS_TECHNICAL_REPORT
Code: https://github.com/joaopauloschuler/less-parameters-llm
- 2 replies
upvoted
an
article
about 1 month ago
Article
DeepSeek-R1 Dissection: Understanding PPO & GRPO Without Any Prior Reinforcement Learning Knowledge
By
β’
β’
70
updated
a
model
about 1 month ago
liked
a
Space
about 1 month ago
Adding Evaluation Results
#1 opened about 1 month ago
by
jlzhou