Deep Fusion Memory adds chunk-level parametric memory to LLMs
With the backbone frozen, training needs 32 times fewer retrieval queries and inference uses no explicit retrieval.
41
AI model
Qwen3-8B-Base
- Mentions · 7d0
- Mentions · 30d1
- All mentions5
- First seenSep 7, 2026
- Last mentioned24d ago
Timeline
5 articles · newest firstKV4-RL: Unlocking Faster Asynchronous Rollouts for RLVR with INT4 KV Cache Quantization
KV4-RL uses selective INT4 KV quantization to speed Qwen3 GRPO rollouts nearly 2x while keeping benchmark accuracy.
52
Everything in Moderation: Per-Domain Coverage Optima and Alignment-Resistant Domain Gaps in Multi-Domain Mid-Training
Controlled mid-training experiments on Qwen3-8B-Base find each domain has a 10-40% coverage optimum and domain gaps survive alignment SFT.
30
ActReview: Rebuttal-Guided Training Data and Rubric Rewards for Actionable Peer Review Generation
ActReview post-trains Qwen3-8B-Base on 40K rebuttal-derived instances with rubric rewards to generate actionable, grounded peer-review feedback, plus a 1,000-instance benchmark.
31
ActReview: Rebuttal-Guided Training Data and Rubric Rewards for Actionable Peer Review Generation
ActReview post-trains Qwen3-8B-Base on OpenReview rebuttals to generate actionable peer-review feedback with grounded revision suggestions, benchmarked on 1,000 curated instances.
30
Appears with
- Qwen3-4B-Baseai model · 2
- KOR-Benchorganization · 1
- OpenRevieworganization · 1
- Mistral-7B-v0.3ai model · 1
- ActReviewai model · 1
- FineWeb-Eduorganization · 1
- OpenReviewproduct · 1
- ActReviewproduct · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.