SITE applies soft gradient-based head selection to inject ICL-derived task embeddings, outperforming prior embedding adaptation and few-shot ICL across generation, reasoning, and NLU tasks on 12 LLMs from 4B to 70B parameters.
Retrieval or global context understanding? on many-shot in-context learning for long-context evaluation
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
representative citing papers
Frontier LLMs with in-context learning and CAS integration solve most algorithmic tasks in theoretical physics when supplied with worked examples.
A survey categorizing scaling in LLM reasoning across input size, steps, rounds, training, and future directions, noting that scaling can negatively affect performance.
citing papers explorer
-
Soft Head Selection for Injecting ICL-Derived Task Embeddings
SITE applies soft gradient-based head selection to inject ICL-derived task embeddings, outperforming prior embedding adaptation and few-shot ICL across generation, reasoning, and NLU tasks on 12 LLMs from 4B to 70B parameters.
-
LLMs with in-context learning for Algorithmic Theoretical Physics
Frontier LLMs with in-context learning and CAS integration solve most algorithmic tasks in theoretical physics when supplied with worked examples.
-
A Survey of Scaling in Large Language Model Reasoning
A survey categorizing scaling in LLM reasoning across input size, steps, rounds, training, and future directions, noting that scaling can negatively affect performance.