Papers with difficult tasks
Adaptive Layer Selection for Layer-Wise Token Pruning in LLM Inference (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated remarkable capabilities in processing long contexts. |
| Approach: | They propose a training-free method that adaptively chooses the selection layer for KV cache reduction . they exploit the variance of token ranks ordered by attention score to optimize decoding . |
| Outcome: | The proposed method outperforms state-of-the-art token pruning methods on InfiniteBench, RULER, and NIAH benchmarks. |
Self-Regulated Sample Diversity in Large Language Models (2024.findings-naacl)
Copied to clipboard
Mingyue Liu, Jonathan Frawley, Sarah Wyer, Hubert P. H. Shum, Sara Uckelman, Sue Black, Chris Willcocks
| Challenge: | Existing methods that require expensive setups or maintain static values during inference are inflexible and require expensive training. |
| Approach: | They propose a self-regulating approach that adjusts sample diversity parameters dynamically based on the input prompt. |
| Outcome: | The proposed method significantly improves the quality of responses generically without model retraining or fine-tuning. |
Multi-Agent Language Learning: Symbolic Mapping (2023.findings-acl)
Copied to clipboard
| Challenge: | Recent work has focused on the emergence of language in cooperative tasks where neural network agents learn a communication protocol from scratch to solve problems together. |
| Approach: | They propose a task transfer method and symbolic mapping architecture to help agents learn a compositional and symmetric language in dialog games. |
| Outcome: | The proposed method can help agents learn a compositional and symmetric language in complex settings like dialog games and the proposed architecture promotes vocabulary expansion. |
Do Language Models Mirror Human Confidence? Exploring Psychological Insights to Address Overconfidence in LLMs (2025.findings-acl)
Copied to clipboard
| Challenge: | Psychology research has shown that humans are poor at estimating their performance on tasks, tending towards underconfidence on easy tasks and overconfidence on difficult tasks. |
| Approach: | They propose to use a self-assessment method to assess confidence in large language models (LLMs) they propose to ask for the answer separately and then use them to improve their accuracy. |
| Outcome: | The proposed method improves confidence calibration and interpretability in QA tasks with different personas. |