proomt

Search

Search posts, papers, and topics

All posts

Hugging Face Daily PapersRuishuo Chen, Xun Wang, Yu Chen1 min readpaperadvanced

The Router Within: Eliciting Native Skill Routing from a Frozen LLM

Read on arxiv.org

Summary

The paper introduces Gavel, a method that extracts a frozen LLM's internal routing signal via two trained linear maps, eliminating the need to embed skill descriptions in the prompt. Experiments on Qwen3‑32B show up to 13.4‑point improvements on task benchmarks and higher skill‑use accuracy compared to larger retrieval‑based systems.

  • Gavel uses a "glance" step: projects mid‑layer activations of the task and each skill through two linear maps to score a full skill library.
  • A "verdict" step resumes forward passes for top‑k skills, combining the model's own likelihood with the glance score as a product of experts.
  • Only the two linear maps are trained; the backbone LLM remains frozen, and per‑skill activation banks are built once at installation.
  • Zero‑shot transfer to three public benchmarks and a new 372‑trajectory SkillTraj benchmark outperforms retrieval‑and‑rerank pipelines that add up to 16 B external parameters.

LLM‑agent developers can achieve accurate skill selection without inflating model size or context, saving compute and simplifying deployment.

8/10
llmretrievalskill routingagents

Related reading

  1. Beyond Top-k Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents

    Beyond Top‑k Skill Retrieval: Diversity‑Aware Skill Routing (DSR) applies a Determinantal Point Process with a query‑residual diversity kernel to rerank skill candidates, balancing relevance and redundancy. On the SkillRouter benchmark it raises recall and full‑coverage, especially for multi‑skill queries, showing that skill routing benefits from set‑selection rather than independent ranking.

    Hugging Face Daily Papersarxiv.org1 minpaper

  3. Learning to solve hard problems in RL for LLMs by never giving up

    The post introduces the *Matthew Effect* in RL‑fine‑tuning of LLMs—performance gains concentrate on tasks the model already solves— and proposes *Never Give Up* (NGU), an adaptive sampling scheme that uses a small k for easy prompts and retries hard prompts with a high‑probability “never give up” loop. Experiments on math (AIME, GSM8k), code (Manufactoria), and larger‑scale setups (DeepScaler) sh…

    Hacker News front pagegithub.io11 minHN1179

  4. Online Learning with LLM Experts from Limited Feedback

    The paper models prompt routing to multiple LLM experts as a bandit problem with limited feedback and proposes algorithms that achieve sublinear regret in both full‑information and bandit settings. Experiments demonstrate that the methods learn effective routing strategies across diverse LLMs using only a small feedback budget.

    Hugging Face Daily Papersarxiv.org2 minpaper

  5. CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents

    CERA-MoA proposes a reinforcement‑learning loop where a router and a set of LLM agents are trained together. A “familiarity” estimator reads mid‑layer hidden states to predict each agent’s competence on a query, letting the router activate only a minimal subset of agents that meet a cumulative confidence threshold. The system also feeds targeted training examples to agents based on their evolving…

    Hugging Face Daily Papersarxiv.org1 minpaper

  6. Article: Your Next DSL Author Is a Language Model

    Typed Domain Grounding (TDG) embeds a DSL inside a mainstream language the LLM already knows (e.g., Kotlin) and uses the host compiler as an oracle. The author describes five building blocks—embedding, choosing a host language with high training‑data frequency, compiler‑driven type safety, a generate‑compile‑repair loop, and an on‑demand teaching tool—and shows measured results from kUML, a Kotli…

    InfoQinfoq.com18 min