š Open to Work
Garrosh Icecream
GarroshIcecream
AI & ML interests
From tiny SLMs to massive LLMs. Iām all about text-to-text fun.
Recent Activity
liked a Space 12 days ago
suvadityamuk/3d-representations-guide liked a model 4 months ago
MisoLabs/MisoTTS updated a collection 9 months ago
Awesome papersOrganizations
None yet
READ ON TOILET
-
Speed Always Wins: A Survey on Efficient Architectures for Large Language Models
Paper ā¢ 2508.09834 ā¢ Published ā¢ 53 -
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Paper ā¢ 2509.02547 ā¢ Published ā¢ 238 -
DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search
Paper ā¢ 2509.25454 ā¢ Published ā¢ 147 -
DeMo: Decoupled Momentum Optimization
Paper ā¢ 2411.19870 ā¢ Published ā¢ 6
Awesome papers
-
Yume: An Interactive World Generation Model
Paper ā¢ 2507.17744 ā¢ Published ā¢ 92 -
SSRL: Self-Search Reinforcement Learning
Paper ā¢ 2508.10874 ā¢ Published ā¢ 97 -
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
Paper ā¢ 2506.06941 ā¢ Published ā¢ 17 -
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper ā¢ 2506.01939 ā¢ Published ā¢ 190
P(DOOM) = 1.0
-
FutureX: An Advanced Live Benchmark for LLM Agents in Future Prediction
Paper ā¢ 2508.11987 ā¢ Published ā¢ 72 -
InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
Paper ā¢ 2508.18265 ā¢ Published ā¢ 221 -
Less is More: Recursive Reasoning with Tiny Networks
Paper ā¢ 2510.04871 ā¢ Published ā¢ 520
English? __Pfff__
Article
P(DOOM) = 1.0
-
FutureX: An Advanced Live Benchmark for LLM Agents in Future Prediction
Paper ā¢ 2508.11987 ā¢ Published ā¢ 72 -
InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
Paper ā¢ 2508.18265 ā¢ Published ā¢ 221 -
Less is More: Recursive Reasoning with Tiny Networks
Paper ā¢ 2510.04871 ā¢ Published ā¢ 520
READ ON TOILET
-
Speed Always Wins: A Survey on Efficient Architectures for Large Language Models
Paper ā¢ 2508.09834 ā¢ Published ā¢ 53 -
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Paper ā¢ 2509.02547 ā¢ Published ā¢ 238 -
DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search
Paper ā¢ 2509.25454 ā¢ Published ā¢ 147 -
DeMo: Decoupled Momentum Optimization
Paper ā¢ 2411.19870 ā¢ Published ā¢ 6
English? __Pfff__
Awesome papers
-
Yume: An Interactive World Generation Model
Paper ā¢ 2507.17744 ā¢ Published ā¢ 92 -
SSRL: Self-Search Reinforcement Learning
Paper ā¢ 2508.10874 ā¢ Published ā¢ 97 -
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
Paper ā¢ 2506.06941 ā¢ Published ā¢ 17 -
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper ā¢ 2506.01939 ā¢ Published ā¢ 190