In-context learning binds model outputs to the demonstrated label tokens as an exhaustive vocabulary, overriding semantic plausibility and causing fixation even with homogeneous or nonsense labels.
arXiv preprint arXiv:2303.03846 , year=
7 Pith papers cite this work. Polarity classification is still indexing.
representative citing papers
In a controlled synthetic setting, transformers implement in-distribution task inference via convex combinations of task vectors and out-of-distribution inference via nearly orthogonal extrapolative representations.
Fine-tuning shows higher proficiency than in-context learning on in-distribution generalization in formal languages, with equal out-of-distribution performance and diverging inductive biases at high proficiency.
Contextual entrainment decreases for semantic contexts but increases for non-semantic ones as LLMs scale, following power-law trends with 4x better resistance to misinformation but 2x more copying of arbitrary tokens.
OPT-BENCH and OPT-Agent evaluate LLM self-optimization in large search spaces, showing stronger models improve via feedback but stay constrained by base capacity and below human performance.
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
Training-free methods for LLM trustworthiness show inconsistent results across dimensions, with clear trade-offs in utility, robustness, and overhead depending on where they intervene during inference.
citing papers explorer
-
In-Context Fixation: When Demonstrated Labels Override Semantics in Few-Shot Classification
In-context learning binds model outputs to the demonstrated label tokens as an exhaustive vocabulary, overriding semantic plausibility and causing fixation even with homogeneous or nonsense labels.
-
Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers
In a controlled synthetic setting, transformers implement in-distribution task inference via convex combinations of task vectors and out-of-distribution inference via nearly orthogonal extrapolative representations.
-
Fine-tuning vs. In-context Learning in Large Language Models: A Formal Language Learning Perspective
Fine-tuning shows higher proficiency than in-context learning on in-distribution generalization in formal languages, with equal out-of-distribution performance and diverging inductive biases at high proficiency.
-
Better and Worse with Scale: How Contextual Entrainment Diverges with Model Size
Contextual entrainment decreases for semantic contexts but increases for non-semantic ones as LLMs scale, following power-law trends with 4x better resistance to misinformation but 2x more copying of arbitrary tokens.
-
OPT-BENCH: Evaluating the Iterative Self-Optimization of LLM Agents in Large-Scale Search Spaces
OPT-BENCH and OPT-Agent evaluate LLM self-optimization in large search spaces, showing stronger models improve via feedback but stay constrained by base capacity and below human performance.
-
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
-
A Systematic Study of Training-Free Methods for Trustworthy Large Language Models
Training-free methods for LLM trustworthiness show inconsistent results across dimensions, with clear trade-offs in utility, robustness, and overhead depending on where they intervene during inference.