Y
Hacker News
new
|
ask
|
show
|
jobs
user:
veryluckyxyz
created:
2014-09-30
karma:
549
submissions:
Scaling Laws for Agent Harnesses via Effective Feedback Compute
1 points
|
0 comments
0 points
|
0 comments
0 points
|
0 comments
Generalizing Test-Time Compute-Optimal Scaling as an Optimizable Graph
2 points
|
0 comments
Hidden drivers of HRM's performance on ARC-AGI
31 points
|
2 comments
Set Block Decoding Is a Language Model Inference Accelerator
4 points
|
0 comments
Deep Think with Confidence
1 points
|
0 comments
A Batch Size and Token NUM- BER Agnostic Learning Rate Scheduler
2 points
|
0 comments
Easily Understand Rdma Technology
1 points
|
1 comments
Model Merging in Pre-Training of Large Language Models
2 points
|
0 comments
Understanding Perception and Reasoning Through Model Merging
2 points
|
0 comments
Building and better understanding vision-language models (2024)
2 points
|
0 comments
HF smolagents computer-agent demo
1 points
|
0 comments
Do Reasoning Models Show Better Verbalized Calibration?
2 points
|
0 comments
Robustly identifying concepts introduced during chat fine-tuning with crosscoder
6 points
|
0 comments
Retrieval with Learned Similarities
3 points
|
0 comments
The Curse of Depth in Large Language Models
1 points
|
0 comments
0 points
|
0 comments
Looking Back at Speculative Decoding
36 points
|
5 comments
Long-Context GRPO
60 points
|
22 comments
HippoRAG: Neurobiologically Inspired Long-Term Memory for LLMs (2024)
65 points
|
4 comments
Learning to Plan and Reason for Evaluation with Thinking-LLM-as-a-Judge
1 points
|
0 comments
Process Reinforcement Through Implicit Rewards
1 points
|
0 comments
Explaining Large Language Models Decisions Using Shapley Values
89 points
|
19 comments
Phi-4 Technical Report
2 points
|
0 comments
Alignment Faking in LLMs [pdf]
2 points
|
1 comments
What Makes Rotary Positional Encodings Useful?
1 points
|
0 comments
Rethinking Softmax: Self-Attention with Polynomial Activations
2 points
|
0 comments
Post-Training Layer Scaling Prevents Forgetting and Enhances Model Merging
1 points
|
0 comments