9. Artificial Hippocampus Networks Adds a fixed-size recurrent memory to sliding-window Transformers, compressing evicted KV into RNN-like states (Mamba2/DN/GDN) trained via self-distillation for long-context efficiency with constant cache and near-linear compute.
LLMS
-

Mem-Agent: 4B LLM with Persistent Memory via GSPO
By
–
8. mem-agent mem-agent is a 4B-parameter LLM trained with GSPO reinforcement learning to develop persistent memory using a scaffold of Python tools and markdown files.
-

ArcMemo: Memory Framework for Abstract Reasoning Tasks
By
–
7. Abstract Reasoning Composition UC San Diego and UMD propose ArcMemo, a test-time memory framework that distills reusable concepts from solution traces, stores them in natural language, and retrieves a relevant subset on future queries.
-
Markovian Thinker: RL Environment for LLM State Management
By
–
6. The Markovian Thinker
— DAIR.AI (@dair_ai) 12 octobre 2025
A new RL thinking environment that keeps an LLM’s effective state constant by chunking long chains of thought and carrying over only a short textual state between chunks.https://t.co/mqXfF1XG6h6. The Markovian Thinker A new RL thinking environment that keeps an LLM’s effective state constant by chunking long chains of thought and carrying over only a short textual state between chunks.
-

Scaling Long-Horizon Reasoning Without Step Labels
By
–
5. Reasoning over Longer Horizons via RL The authors show that you can scale long-horizon reasoning without step labels or heavy scaffolding.
-

Inoculation Prompting: Safety Technique for Flawed Training Data
By
–
4. Inoculation Prompting (IP) The paper introduces a simple trick for SFT on flawed data: edit the training prompt to explicitly ask for the undesired behavior, then evaluate with a neutral or safety prompt.
-

Agentic Context Engineering: Modular Framework for LLM Playbooks
By
–
3. Agentic Context Engineering (ACE) Presents a modular context-engineering framework that grows and refines an LLM’s working context like a playbook, not a terse prompt.
-
Agent responses feel more reliable than personal Gemini
By
–
I am still playing with Agent responses, which are powered by the same Gemini 2.5 Pro, but the output feels more reliable than what we have on personal Gemini. Full scoop
-
Qwen 3 release questioned technical report limitations
By
–
I think the only one I might question the significance of in your list is Qwen 3. They didn't even release a base model. The technical report was quite disappointing as well.
-
vLLM Limitations in Broader AI Landscape
By
–
vLLM is a small part of the whole picture. They're massively lacking.