Tag feed

#reinforcement-learning

255 posts54 followers

Explore Hashnode

Alternatives

Trending tags this week

ARAleksei Romanovgfactor.hashnode.dev4h ago · 15 min read

Multi-Reward RL, Part 3: GDPO + CISPO + REPO-R at 27B, and the Advantage Floor That Stopped Learning

Part 1 explained how PPO, GRPO, DAPO and GDPO turn several rewards into one learning signal. Part 2 benchmarked the trainers on a 14B model for 50 steps and found a trap: CISPO won the training curves

0

NONick One-Dropnickonedrop.hashnode.dev2d ago · 3 min read

Multi-agent cooperation through in-context co-player inference (2026)

This paper takes a decentralized multi-agent reinforcement learning approach. What does that mean here? For context, let’s look at another DeepMind paper with overlapping authors: Multi-agent cooperat

0

ARAleksei Romanovgfactor.hashnode.dev3d ago · 14 min read

Multi-Reward RL, Part 2: Benchmarking GRPO, DAPO, and CISPO on Unseen Tasks

Follow-up: Part 3 scales the CISPO + REPO-R recipe to Qwen3.8-27B and 600 steps, with a one-change-per-run holdout ladder and an advantage-floor failure we found in a harsher environment. Part 1 analy

0

ARAleksei Romanovgfactor.hashnode.dev3d ago · 14 min read

The Verifier Design Playbook: How to Build RLVR Gyms That Models Can't Game

In traditional machine learning, your loss function is a clean mathematical equation: mean squared error, cross-entropy, or cosine distance. The math is simple, deterministic, and impossible for the m

0

ARAleksei Romanovgfactor.hashnode.dev5d ago · 18 min read

The Specialized Agent Shift: How Open Models Are Beating Frontier APIs in Production

For the past two years, the standard enterprise AI strategy looked suspiciously like a parlor trick: take an 8,000-token system prompt packed with JSON schemas, API documentation, and polite formattin

0

ARAleksei Romanovgfactor.hashnode.dev6d ago · 15 min read

Long-Horizon Agents: Why Multi-Turn Reasoning Breaks and the Practical Training Tricks That Fix It

Almost every modern language model looks impressive on a two-step demo. You ask it to check a database or summarize a document, it calls the right tool, formats the answer, and looks like an autonomou

0

ARAleksei Romanovgfactor.hashnode.devOct 2 · 13 min read

We Tried ISO-AdamW. AdamW Kept Its Job.

Four answers. After implementing a new optimizer, fixing a painfully slow matrix operation, and running both versions through the same math exam, that was the gap: 758 correct answers for ISO-AdamW ve

0

ARAleksei Romanovgfactor.hashnode.devSep 28 · 11 min read

From RLHF to RLVR: The Evolution of Reward Signals and the Battle Against Reward Hacking

If you want to understand why reinforcement learning in AI is both exhilarating and infuriating, you only need to remember one golden rule: models do not optimize for what you want; they optimize for

0

9S99 sunnysunny99.hashnode.devSep 28 · 9 min read

The agent treats the fence as a bug

Epistemic status: A close reading of the misbehavior and access-control sections of DeepSeek's DSec report, plus a proposed direction drawn from my own runtime work. The incidents are as DeepSeek repo

0

NONick One-Dropnickonedrop.hashnode.devSep 27 · 3 min read

Experiential Reinforcement Learning (2026)

The basic reinforcement learning loop is simple: try something, receive a reward, and repeat. Poor behavior is gradually corrected as reward signals accumulate across trials. This paper proposes Exper

0

#reinforcement-learning

Search Hashnode

#reinforcement-learning

Explore Hashnode

Trending tags this week

Multi-Reward RL, Part 3: GDPO + CISPO + REPO-R at 27B, and the Advantage Floor That Stopped Learning

Multi-agent cooperation through in-context co-player inference (2026)

Multi-Reward RL, Part 2: Benchmarking GRPO, DAPO, and CISPO on Unseen Tasks

The Verifier Design Playbook: How to Build RLVR Gyms That Models Can't Game

The Specialized Agent Shift: How Open Models Are Beating Frontier APIs in Production

Long-Horizon Agents: Why Multi-Turn Reasoning Breaks and the Practical Training Tricks That Fix It

We Tried ISO-AdamW. AdamW Kept Its Job.

From RLHF to RLVR: The Evolution of Reward Signals and the Battle Against Reward Hacking

The agent treats the fence as a bug

Experiential Reinforcement Learning (2026)