Ben Kelly PRO
YellowjacketGames
AI & ML interests
None yet
Recent Activity
updated
a collection
about 3 hours ago
[papers] Gameplay Optimization
updated
a collection
about 14 hours ago
[ papers ] Sports Tech
updated
a collection
about 14 hours ago
[ papers ] Sports Tech
Organizations
Research directly related to Chess technology.
Models that run well on a *standalone* RTX a6000's 48gb of VRAM.
Toaster Tier but not iGPU
TPS can be as low as 1.0, seriously. its SLOW.
-
unsloth/GLM-4.7-GGUF
Text Generation • 358B • Updated • 122k • 186 -
unsloth/DeepSeek-R1-0528-GGUF
Text Generation • 671B • Updated • 4.18k • 193 -
unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF
Image-to-Text • 401B • Updated • 6.24k • 42 -
unsloth/MiniMax-M2.1-GGUF
Text Generation • 229B • Updated • 137k • 159
-
Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment
Paper • 2601.14249 • Published • 8 -
Fiddler: CPU-GPU Orchestration for Fast Inference of Mixture-of-Experts Models
Paper • 2402.07033 • Published • 19 -
MeepleLM: A Virtual Playtester Simulating Diverse Subjective Experiences
Paper • 2601.07251 • Published • 8 -
GameTalk: Training LLMs for Strategic Conversation
Paper • 2601.16276 • Published • 10
-
A BERTology View of LLM Orchestrations: Token- and Layer-Selective Probes for Efficient Single-Pass Classification
Paper • 2601.13288 • Published • 12 -
nvidia/Nemotron-Orchestrator-8B
Text Generation • 8B • Updated • 18.7k • 532 -
OptiMind: Teaching LLMs to Think Like Optimization Experts
Paper • 2509.22979 • Published • 3 -
OpenDevin: An Open Platform for AI Software Developers as Generalist Agents
Paper • 2407.16741 • Published • 75
Research papers that may contribute to a broader approach to teaching machines how to play complex strategy games beyond just Chess.
-
OptiMind: Teaching LLMs to Think Like Optimization Experts
Paper • 2509.22979 • Published • 3 -
LFM2 Technical Report
Paper • 2511.23404 • Published • 51 -
Zero-Overhead Introspection for Adaptive Test-Time Compute
Paper • 2512.01457 • Published • 1 -
Confidence Estimation for LLMs in Multi-turn Interactions
Paper • 2601.02179 • Published • 16
The best little card under 100 euros. Full Precision vs Quants not benchmarked. This card is so much better at running inference than you realize.
hey, you gotta try. any precision acceptable here, QA check the actual result quality. at your own risk, fool.
The stuff we actually use, pruned on an ongoing basis.
-
Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment
Paper • 2601.14249 • Published • 8 -
Fiddler: CPU-GPU Orchestration for Fast Inference of Mixture-of-Experts Models
Paper • 2402.07033 • Published • 19 -
MeepleLM: A Virtual Playtester Simulating Diverse Subjective Experiences
Paper • 2601.07251 • Published • 8 -
GameTalk: Training LLMs for Strategic Conversation
Paper • 2601.16276 • Published • 10
-
A BERTology View of LLM Orchestrations: Token- and Layer-Selective Probes for Efficient Single-Pass Classification
Paper • 2601.13288 • Published • 12 -
nvidia/Nemotron-Orchestrator-8B
Text Generation • 8B • Updated • 18.7k • 532 -
OptiMind: Teaching LLMs to Think Like Optimization Experts
Paper • 2509.22979 • Published • 3 -
OpenDevin: An Open Platform for AI Software Developers as Generalist Agents
Paper • 2407.16741 • Published • 75
Research directly related to Chess technology.
Research papers that may contribute to a broader approach to teaching machines how to play complex strategy games beyond just Chess.
-
OptiMind: Teaching LLMs to Think Like Optimization Experts
Paper • 2509.22979 • Published • 3 -
LFM2 Technical Report
Paper • 2511.23404 • Published • 51 -
Zero-Overhead Introspection for Adaptive Test-Time Compute
Paper • 2512.01457 • Published • 1 -
Confidence Estimation for LLMs in Multi-turn Interactions
Paper • 2601.02179 • Published • 16
Models that run well on a *standalone* RTX a6000's 48gb of VRAM.
The best little card under 100 euros. Full Precision vs Quants not benchmarked. This card is so much better at running inference than you realize.
Toaster Tier but not iGPU
hey, you gotta try. any precision acceptable here, QA check the actual result quality. at your own risk, fool.
TPS can be as low as 1.0, seriously. its SLOW.
-
unsloth/GLM-4.7-GGUF
Text Generation • 358B • Updated • 122k • 186 -
unsloth/DeepSeek-R1-0528-GGUF
Text Generation • 671B • Updated • 4.18k • 193 -
unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF
Image-to-Text • 401B • Updated • 6.24k • 42 -
unsloth/MiniMax-M2.1-GGUF
Text Generation • 229B • Updated • 137k • 159
The stuff we actually use, pruned on an ongoing basis.