slabhead/ml-papers-mix-2026-04

ML Research Paper Mix — 50 Papers

A curated cross-section of recent Machine Learning research pulled via Hugging Face paper search on 2026-04-24. Organized into six thematic buckets so a reader can scan for alignment with their subfield.

Each paper links to hf.co/papers/<arxiv-id>, which mirrors the arXiv abstract and aggregates community discussion. No PDFs are stored here — this is a pointer index.

---

1. LLM Reasoning, Agents & Tool Use (10)

#	Paper	Date	Link
1	Rethinking the Role of Entropy in Optimizing Tool-Use Behaviors for LLM Agents	2026-02-02	https://hf.co/papers/2602.02050
2	MCP-RADAR: Multi-Dimensional Benchmark for Evaluating Tool Use in LLMs	2025-05-22	https://hf.co/papers/2505.16700
3	From Correction to Mastery: Reinforced Distillation of LLM Agents	2025-09-12	https://hf.co/papers/2509.14257
4	ToolMind: A Large-Scale, Reasoning-Enhanced Tool-Use Dataset	2025-11-12	https://hf.co/papers/2511.15718
5	SE-Agent: Self-Evolution Trajectory Optimization in Multi-Step Reasoning	2025-08-04	https://hf.co/papers/2508.02085
6	SMART: Self-Aware Agent for Tool Overuse Mitigation	2025-02-17	https://hf.co/papers/2502.11435
7	Git Context Controller: Manage the Context of LLM-based Agents like Git	2025-07-30	https://hf.co/papers/2508.00031
8	Information Gain-based Policy Optimization for Multi-Turn LLM Agents	2025-10-16	https://hf.co/papers/2510.14967
9	DualSpec: Accelerating Deep Research Agents via Dual-Process Action Speculation	2026-03-08	https://hf.co/papers/2603.07416
10	Mozi: Governed Autonomy for Drug Discovery LLM Agents	2026-03-04	https://hf.co/papers/2603.03655

2. Diffusion Models & Generative Video (8)

#	Paper	Date	Link
11	Toward Lightweight and Fast Decoders for Diffusion Models	2025-03-06	https://hf.co/papers/2503.04871
12	GenTron: Delving Deep into Diffusion Transformers for Image and Video	2023-12-07	https://hf.co/papers/2312.04557
13	DreamVideo: High-Fidelity Image-to-Video Generation	2023-12-05	https://hf.co/papers/2312.03018
14	H3AE: High Compression, High Speed, High Quality AutoEncoder for Video Diffusion	2025-04-14	https://hf.co/papers/2504.10567
15	Language Model Beats Diffusion — Tokenizer is Key to Visual Generation	2023-10-09	https://hf.co/papers/2310.05737
16	Importance-based Token Merging for Diffusion Models	2024-11-23	https://hf.co/papers/2411.16720
17	Scalable Adaptive Computation for Iterative Generation (RIN)	2022-12-22	https://hf.co/papers/2212.11972
18	Taming Rectified Flow for Inversion and Editing	2024-11-07	https://hf.co/papers/2411.04746

3. RLHF & Alignment (8)

#	Paper	Date	Link
19	Balanced Actor Initialization: Stable RLHF Training of Distillation-Based Reasoning Models	2025-08-30	https://hf.co/papers/2509.00309
20	StatLLaMA: Multi-Stage Training for Domain-Optimized Statistical LLMs	2025-12-26	https://hf.co/papers/2601.09718
21	Improving RLHF with Efficient Reward Model Ensemble	2024-01-30	https://hf.co/papers/2401.16635
22	RLTHF: Targeted Human Feedback for LLM Alignment	2025-02-19	https://hf.co/papers/2502.13417
23	Position: The Complexity of Perfect AI Alignment — Formalizing the RLHF Trilemma	2025-11-23	https://hf.co/papers/2511.19504
24	Towards Reliable Alignment: Uncertainty-aware RLHF	2024-10-31	https://hf.co/papers/2410.23726
25	Pairwise Proximal Policy Optimization: Harnessing Relative Feedback for LLM Alignment	2023-09-30	https://hf.co/papers/2310.00212
26	Mitigating the Alignment Tax of RLHF	2023-09-12	https://hf.co/papers/2309.06256

4. Mechanistic Interpretability (8)

#	Paper	Date	Link
27	Circuit Insights: Towards Interpretability Beyond Activations	2025-10-16	https://hf.co/papers/2510.14936
28	Automated Circuit Interpretation via Probe Prompting	2025-11-10	https://hf.co/papers/2511.07002
29	The Local Interaction Basis: Identifying Computationally-Relevant and Sparsely Interacting Features	2024-05-17	https://hf.co/papers/2405.10928
30	Causally Grounded Mechanistic Interpretability for LLMs with Faithful Natural-Language Explanations	2026-02-13	https://hf.co/papers/2603.09988
31	From Neurons to Neutrons: A Case Study in Interpretability	2024-05-27	https://hf.co/papers/2405.17425
32	Hierarchical Sparse Circuit Extraction from Billion-Parameter Language Models	2026-01-19	https://hf.co/papers/2601.12879
33	Transcoders Find Interpretable LLM Feature Circuits	2024-06-17	https://hf.co/papers/2406.11944
34	Automatically Identifying Local and Global Circuits with Linear Computation Graphs	2024-05-22	https://hf.co/papers/2405.13868

5. Multimodal & Vision-Language Models (8)

#	Paper	Date	Link
35	Xmodel-VLM: A Simple Baseline for Multimodal Vision Language Model	2024-05-15	https://hf.co/papers/2405.09215
36	UniFusion: Vision-Language Model as Unified Encoder in Image Generation	2025-10-14	https://hf.co/papers/2510.12789
37	V-Agent: An Interactive Video Search System Using Vision-Language Models	2025-11-04	https://hf.co/papers/2512.16925
38	HyCodePolicy: Hybrid Language Controllers for Multimodal Monitoring in Embodied Agents	2025-08-04	https://hf.co/papers/2508.02629
39	ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Models	2026-01-16	https://hf.co/papers/2601.11404
40	Query-Kontext: A Unified Multimodal Model for Image Generation and Editing	2025-09-30	https://hf.co/papers/2509.26641
41	BLIP-FusePPO: A Vision-Language DRL Framework for Lane Keeping	2025-10-25	https://hf.co/papers/2510.22370
42	DuetSVG: Unified Multimodal SVG Generation with Internal Visual Guidance	2025-12-11	https://hf.co/papers/2512.10894

6. Efficient Inference: Quantization, Speculative Decoding, MoE (8)

#	Paper	Date	Link
43	Speculative Decoding Meets Quantization: Compatibility Evaluation + Hierarchical Framework	2025-05-28	https://hf.co/papers/2505.22179
44	MoE-Spec: Expert Budgeting for Efficient Speculative Decoding	2026-02-17	https://hf.co/papers/2602.16052
45	Speculative MoE: Communication-Efficient Parallel MoE Inference	2025-03-06	https://hf.co/papers/2503.04398
46	Accelerating Speculative Decoding using Dynamic Speculation Length (DISCO)	2024-05-07	https://hf.co/papers/2405.04304
47	MoE-SpAc: Efficient MoE Inference Based on Speculative Activation Utility	2026-02-12	https://hf.co/papers/2603.09983
48	Jakiro: Boosting Speculative Decoding with Decoupled Multi-Head via MoE	2025-02-10	https://hf.co/papers/2502.06282
49	Turning Trash into Treasure: Accelerating Inference with Token Recycling	2024-08-16	https://hf.co/papers/2408.08696
50	Speculative Decoding Scaling Laws (SDSL): Throughput Optimization Made Simple	2026-02-25	https://hf.co/papers/2603.11053

---

Source: Hugging Face Papers via paper_search MCP — 2026-04-24. Curation method: 6 thematic semantic queries × 8–10 top results each.