Copertina

One Task Vector is not Enough: A Large-Scale Study for In-Context Learning

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Tikhonov, Pavel, Oseledets, Ivan, Tutubalina, Elena
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

I Have Covered All the Bases Here: Interpreting Reasoning Features in Large Language Models via Sparse Autoencoders
di: Galichin, Andrey, et al.
Pubblicazione: (2025)

Team Anotheroption at SemEval-2025 Task 8: Bridging the Gap Between Open-Source and Proprietary LLMs in Table QA
di: Evkarpidi, Nikolas, et al.
Pubblicazione: (2025)

Exploring the Hidden Capacity of LLMs for One-Step Text Generation
di: Mezentsev, Gleb, et al.
Pubblicazione: (2025)

Confidence Estimation for Error Detection in Text-to-SQL Systems
di: Somov, Oleg, et al.
Pubblicazione: (2025)

Geopolitical biases in LLMs: what are the "good" and the "bad" countries according to contemporary language models
di: Salnikov, Mikhail, et al.
Pubblicazione: (2025)

When Punctuation Matters: A Large-Scale Comparison of Prompt Robustness Methods for LLMs
di: Seleznyov, Mikhail, et al.
Pubblicazione: (2025)

Label Words as Local Task Vectors in In-Context Learning
di: Zheng, Bowen, et al.
Pubblicazione: (2024)

CoRoVA: Compressed Representations for Vector-Augmented Code Completion
di: Cherniuk, Daria, et al.
Pubblicazione: (2025)

CLEAR: Character Unlearning in Textual and Visual Modalities
di: Dontsov, Alexey, et al.
Pubblicazione: (2024)

SumHiS: Extractive Summarization Exploiting Hidden Structure
di: Pavel, Tikhonov, et al.
Pubblicazione: (2024)

Diagonal Batching Unlocks Parallelism in Recurrent Memory Transformers for Long Contexts
di: Sivtsov, Danil, et al.
Pubblicazione: (2025)

Jailbreaking? One Step Is Enough!
di: Zheng, Weixiong, et al.
Pubblicazione: (2024)

Anatomy of Unlearning: The Dual Impact of Fact Salience and Model Fine-Tuning
di: Borisiuk, Anna, et al.
Pubblicazione: (2026)

DeCoVec: Building Decoding Space based Task Vector for Large Language Models via In-Context Learning
di: Li, Feiyang, et al.
Pubblicazione: (2026)

On the Spatial Structure of Mixture-of-Experts in Transformers
di: Bershatsky, Daniel, et al.
Pubblicazione: (2025)

BALI: Enhancing Biomedical Language Representations through Knowledge Graph and Language Model Alignment
di: Sakhovskiy, Andrey, et al.
Pubblicazione: (2025)

Language Models for Text Classification: Is In-Context Learning Enough?
di: Edwards, Aleksandra, et al.
Pubblicazione: (2024)

Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning
di: Yang, Haolin, et al.
Pubblicazione: (2025)

Distributional Alignment as a Criterion for Designing Task Vectors in In-Context Learning
di: Kwon, Jihoon, et al.
Pubblicazione: (2026)

Back to Basics: Revisiting Exploration in Reinforcement Learning for LLM Reasoning via Generative Probabilities
di: Li, Pengyi, et al.
Pubblicazione: (2026)

The Chronicles of RiDiC: Generating Datasets with Controlled Popularity Distribution for Long-form Factuality Evaluation
di: Braslavski, Pavel, et al.
Pubblicazione: (2026)

Context is Enough: Empirical Validation of $\textit{Sequentiality}$ on Essays
di: Sunny, Amal, et al.
Pubblicazione: (2025)

SynthDetoxM: Modern LLMs are Few-Shot Parallel Detoxification Data Annotators
di: Moskovskiy, Daniil, et al.
Pubblicazione: (2025)

Emergence and Effectiveness of Task Vectors in In-Context Learning: An Encoder Decoder Perspective
di: Han, Seungwook, et al.
Pubblicazione: (2024)

Induction Signatures Are Not Enough: A Matched-Compute Study of Load-Bearing Structure in In-Context Learning
di: Sabry, Mohammed, et al.
Pubblicazione: (2025)

Task Prompt Vectors: Effective Initialization through Multi-Task Soft-Prompt Transfer
di: Belanec, Robert, et al.
Pubblicazione: (2024)

Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs
di: Afonin, Nikita, et al.
Pubblicazione: (2025)

One Word Is Not Enough: Simple Prompts Improve Word Embeddings
di: Ranjan, Rajeev
Pubblicazione: (2025)

Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning
di: Huang, Brandon, et al.
Pubblicazione: (2024)

Pitfalls of Scale: Investigating the Inverse Task of Redefinition in Large Language Models
di: Stringli, Elena, et al.
Pubblicazione: (2025)

LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformers
di: Razzhigaev, Anton, et al.
Pubblicazione: (2025)

Humor Mechanics: Advancing Humor Generation with Multistep Reasoning
di: Tikhonov, Alexey, et al.
Pubblicazione: (2024)

One Trigger Token Is Enough: A Defense Strategy for Balancing Safety and Usability in Large Language Models
di: Gu, Haoran, et al.
Pubblicazione: (2025)

Position: Enough of Scaling LLMs! Lets Focus on Downscaling
di: Goel, Yash, et al.
Pubblicazione: (2025)

Adaptive Task Vectors for Large Language Models
di: Kang, Joonseong, et al.
Pubblicazione: (2025)

Does In-Context Learning Really Learn? Rethinking How Large Language Models Respond and Solve Tasks via In-Context Learning
di: Long, Quanyu, et al.
Pubblicazione: (2024)

Causal Interventions on Continuous Variables: A Case Study on Verb Bias in Steering Vectors for In-Context Learning
di: Zhou, Zhenghao Herbert, et al.
Pubblicazione: (2026)

Branching Narratives: Character Decision Points Detection
di: Tikhonov, Alexey
Pubblicazione: (2024)

Knowledge Graph Representation for Political Information Sources
di: Osmonova, Tinatin, et al.
Pubblicazione: (2024)

A Case Study on Context-Aware Neural Machine Translation with Multi-Task Learning
di: Appicharla, Ramakrishna, et al.
Pubblicazione: (2024)