#benchmark reward hacking
69 related articles
Three Core Gaps in Multimodal LLMs: Fr…
·4 min
Three Core Gaps in Multimodal LLMs: Frontier Research from Microsoft Research India
Microsoft Research India reveals three core gaps in multimodal LLMs: visual perception blindspots, cognitive hallucination, and architectural limitations. Explores Faithful GRPO, behavior modeling, and model alignment breakthroughs.
Read more →
·4 min
RAM: First-Principles Approach to 50x Faster RL Post-Training for Diffusion Models
RAM (Reinforce Adjoint Matching) achieves 50x faster RL post-training for diffusion models by discarding path costs, combining ODE sampling with decorrelated training objectives. A deep dive into RAM's core principles and experiments vs. Flow-GRPO.
Read more →
The 'One-Step Trap' in AI Research: Wh…
·2 min
The 'One-Step Trap' in AI Research: Why Local Optima Lead to Dead Ends
What is the 'One-Step Trap' in AI research? A deep dive into how greedy thinking locks research directions, the limits of incremental improvements, and how multi-step planning and exploration-exploitation balance enable real breakthroughs.
Read more →
·3 min
In-Depth Review of GPT 5.6 Soul: Sub-Agent Systems, Government Review, and Competitor Comparison
In-depth analysis of GPT 5.6 Soul: multi-sub-agent parallel architecture, Ultra Mode coding in practice, the controversy behind its 91.9% Terminal Bench score, and the trend of frontier AI entering government review.
Read more →
·2 min
VersatIL: A Modular PyTorch Framework Tackling Code Fragmentation in Imitation Learning
VersatIL is a modular PyTorch framework for robot imitation learning that decouples data, network architecture, algorithm, and objective. Supports ACT, Diffusion Policy, pi0, and LeRobot format.
Read more →
·6 min
GPT-5.6 Sol Hands-On: How Multi-Agent Parallelism Is Reshaping AI Coding Workflows
Based on Fireship's review, an in-depth look at GPT-5.6 Sol's Ultra Mode multi-agent parallelism, its 91.9% Terminal Bench score, and how it differs from Claude Fable in cost, speed, and precision.
Read more →
·3 min
GPT-5.6 Launches with Three Models: Sol/Terra/Luna and ChatGPT Work Reshape How We Work
OpenAI launches the GPT-5.6 family—Sol, Terra, and Luna—alongside ChatGPT Work, a new desktop app, and Sites. AI evolves from a chat tool into a true productivity partner, completing financial analysis, presentations, and cross-platform collaboration in one operation.
Read more →
·5 min
In-Depth Analysis of the GPT-5.6 EQ Test: Progress and Limitations in Emotional Intelligence
A Reddit user ran EQ tests on ChatGPT 5.5 and 5.6, covering meeting emotion ranking, chess-behavior judgment, and facial attractiveness. Version 5.6 shows clear gains in multimodal emotional understanding, but social common sense remains a core weakness.
Read more →
·5 min
GPT-5.6 In-Depth Review: Crushing Competitors on Value, But Its Jailbreak Vulnerabilities Can't Be Ignored
OpenAI's GPT-5.6 Soul, Terra & Luna are priced at one-third of Claude, leading Anthropic Fable on many benchmarks. We analyze its value, reasoning, and jailbreak risks.
Read more →
·5 min
ChatGPT Work Officially Launches: How GPT-5.6 Is Reshaping the Way You Work
OpenAI launches the GPT-5.6 model family (Sol/Terra/Luna) and ChatGPT Work, enabling automated financial analysis, local file operations, Codex coding, and cross-app workflows—AI officially becomes a real work partner.
Read more →
·5 min
GPT-5.6 Released: Outperforms Rivals but Halted by Government—Cybersecurity Capabilities Trigger Regulatory Red Line
OpenAI releases GPT-5.6 (Sol/Terra/Luna), beating Anthropic on Terminal Bench at ~40% lower cost. But its cybersecurity capabilities hit danger thresholds, limiting access to trusted partners at government request.
Read more →
·6 min
Zuckerberg Admits AI Progress Falls Short of Expectations: The Industry Truth Behind the Layoffs
Meta laid off 8,000 to bet on AI, yet Zuckerberg admits AI agents fell short of expectations. A look at the collective 'AI reflection' among OpenAI, Microsoft, and Google, plus research on AI's selective impact on jobs.
Read more →
·2 min
AI Agents for SGLang Kernel Optimization: Real-World Performance Gains and Engineering Methodology
SGLang's team converted expert knowledge into agent skills, achieving 71.4% throughput gains, TTFT reduced from 456ms to 168ms. A deep dive into agent-assisted kernel optimization methodology.
Read more →
·3 min
GPT-5.6 Released: Sol, Terra, and Luna Reshape AI Programming and Office Automation
OpenAI launches GPT-5.6 with three models — Sol, Terra, and Luna — plus ChatGPT Work, a new desktop app, and Hosted Sites. Codex now autonomously trains models.
Read more →
·6 min
OpenAI Releases GPT-5.6: Cybersecurity Capabilities Take Center Stage
OpenAI launches the GPT-5.6 model family with cybersecurity as its biggest highlight. A deep analysis of GPT-5.6's differentiation, double-edged-sword effect, and enterprise strategy.
Read more →
·6 min
Tencent Hunyuan HY3 Official Version Open-Sourced: The Strongest Under 300B, at 1 Yuan per Million Tokens
Tencent Hunyuan HY3 official version is open-sourced under Apache 2.0, priced as low as 1 yuan per million input tokens, with major gains in agents, reasoning, coding, and long context. On the same day, Meituan open-sourced its trillion-parameter LongCat 2.0.
Read more →
·7 min
Flagship Model vs. Local 27B: Which Follows Code Instructions More Faithfully?
An experiment having Claude Opus and a 27B local open-source model each build a CoD game reveals frontier LLMs' problem of over-inferring intent—Opus added wallhack cheats on its own, while the small local model faithfully followed instructions.
Read more →
LLM Security Benchmarking: Current Sta…
·4 min
LLM Security Benchmarking: Current State, Challenges, and Practical Guide
Why is it so hard to establish unified LLM security benchmarks? This article analyzes core challenges in LLM security evaluation—covering jailbreaks, prompt injection, red teaming, and more—with practical strategies for developers.
Read more →
·4 min
Reinforcement Learning for AI Agents: A Deep Dive into RLHF and Agentic RL
A deep dive into RL for AI agents: from RLHF to Agentic RL, covering PPO vs. GRPO, sparse rewards, tool-calling optimization, and verifiable rewards.
Read more →
·7 min
Andrew Ng's Prompt Engineering Course Essentials: Core Methodology Every Developer Must Learn
An in-depth analysis of the essentials of Andrew Ng and OpenAI's ChatGPT Prompt Engineering course. Covers the difference between base and instruction-tuned models, two core prompting principles, and how to wield LLM APIs to build apps.
Read more →