MiniMaxAI/MiniMax-M2.1
Text Generation
ā¢
229B
ā¢
Updated
ā¢
121
ā¢
ā¢
366
Noob
noobmldude
AI & ML interests
Explainable AI
Recent Activity
liked
a model
about 12 hours ago
MiniMaxAI/MiniMax-M2.1
upvoted
a
paper
1 day ago
Shadow-FT: Tuning Instruct via Base
upvoted
a
collection
1 day ago
Model Merging
Organizations
liked
a
model
about 12 hours ago
upvoted
a
paper
1 day ago
upvoted
a
collection
1 day ago
liked
a
model
8 days ago
liked
a
model
10 days ago
liked
a
dataset
10 days ago
upvoted
a
collection
10 days ago
liked
a
model
12 days ago
liked
a
Space
12 days ago
upvoted
an
article
16 days ago
Article
Tricks from OpenAI gpt-oss YOU š«µ can use with transformers
- +5
ā¢
174
liked
2
models
16 days ago
liked
a
model
18 days ago
liked
a
model
19 days ago
upvoted
a
paper
23 days ago
#2 opened 23 days ago
by
noobmldude
liked
a
Space
23 days ago
liked
a
Space
26 days ago
liked
a
model
about 1 month ago
reacted to
codelion's
post with ā¤ļø
2 months ago
Post
3258
š§ Introducing Ellora Recipe #6: Execution-Aware World Model for Qwen3-4B-Thinking
Teaching LLMs to understand not just what code does, but HOW it executes at runtime!
Inspired by Meta's CWM (Code World Model) research, this LoRA adapter adds execution awareness to Qwen3-4B-Thinking-2507. The model learns to predict variable states, trace program execution step-by-step, and debug code by understanding runtime behavior.
š Key Innovation:
We combine Qwen3's native thinking capabilities with real Python execution traces captured via sys.settrace(). The model is trained using GRPO with a custom reward function that scores execution prediction accuracy.
š Training Approach:
- Hybrid Magpie-style code generation
- Real execution tracing for ground truth
- Self-supervised learning (no manual annotations!)
- 298 training samples with execution traces
āØ What it does:
- Predicts variable states at each line of code
- Explains execution flow with thinking tags
- Helps debug by understanding runtime behavior
- Works as a "neural debugger"
šÆ Results:
- 20% overall accuracy on execution prediction
- 33.3% mean state accuracy
- Trained on Qwen3-4B-Thinking (262K context, 4B params)
š Links:
Model: codelion/Qwen3-4B-execution-world-model-lora
Dataset: codelion/execution-world-model-dataset
GitHub Recipe: https://github.com/codelion/ellora
Notebook: https://github.com/codelion/ellora/blob/main/Ellora_Recipe_6_Execution_World_Model_Thinking_LoRA.ipynb
Part of the Ellora project - standardized LoRA recipes for enhancing LLM capabilities. All recipes use self-supervised data generation and work with existing infrastructure (PEFT, LoRAX, vLLM).
#LLM #LoRA #CodeGeneration #WorldModel #Qwen #AI #MachineLearning
Teaching LLMs to understand not just what code does, but HOW it executes at runtime!
Inspired by Meta's CWM (Code World Model) research, this LoRA adapter adds execution awareness to Qwen3-4B-Thinking-2507. The model learns to predict variable states, trace program execution step-by-step, and debug code by understanding runtime behavior.
š Key Innovation:
We combine Qwen3's native thinking capabilities with real Python execution traces captured via sys.settrace(). The model is trained using GRPO with a custom reward function that scores execution prediction accuracy.
š Training Approach:
- Hybrid Magpie-style code generation
- Real execution tracing for ground truth
- Self-supervised learning (no manual annotations!)
- 298 training samples with execution traces
āØ What it does:
- Predicts variable states at each line of code
- Explains execution flow with thinking tags
- Helps debug by understanding runtime behavior
- Works as a "neural debugger"
šÆ Results:
- 20% overall accuracy on execution prediction
- 33.3% mean state accuracy
- Trained on Qwen3-4B-Thinking (262K context, 4B params)
š Links:
Model: codelion/Qwen3-4B-execution-world-model-lora
Dataset: codelion/execution-world-model-dataset
GitHub Recipe: https://github.com/codelion/ellora
Notebook: https://github.com/codelion/ellora/blob/main/Ellora_Recipe_6_Execution_World_Model_Thinking_LoRA.ipynb
Part of the Ellora project - standardized LoRA recipes for enhancing LLM capabilities. All recipes use self-supervised data generation and work with existing infrastructure (PEFT, LoRAX, vLLM).
#LLM #LoRA #CodeGeneration #WorldModel #Qwen #AI #MachineLearning