No article text was recorded for this version.
synthetic-to-natural generalization gap
Version 1 of 147 papers48 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Constrained Fact Verification for FEVER
- Blog: A PDP Architecture For Processing Sentences With Relative Clauses
- Blog: Language Models can be Deductive Solvers
- Blog: A Percolation Model of Emergence: Analyzing Transformers Trained on a Formal Language
- Blog: In-Context Learning Dynamics with Random Binary Sequences
- Blog: APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning
- Blog: In-context Learning on Function Classes Unveiled for Transformers
- Blog: RelNet: End-to-end Modeling of Entities & Relations
- Blog: Position: Understanding LLMs Requires More Than Statistical Generalization
- Blog: How Do Position Encodings Affect Length Generalization? Case Studies On In-Context Function Learning
- Blog: Logical Parsing from Natural Language Based on a Neural Translation Model
- Blog: LSTMs Exploit Linguistic Attributes of Data
- Blog: Sparse Autoencoders Reveal Temporal Difference Learning in Large Language Models
- Blog: Closing Brackets with Recurrent Neural Networks
- Blog: Representational Homomorphism Error Predicts Compositional Generalization In Language Models
- Blog: Quality over Quantity in Attention Layers: When Adding More Heads Hurts
- Blog: Transformers are uninterpretable with myopic methods: a case study with bounded Dyck grammars
- Blog: Towards AI-Complete Question Answering: A Set of Prerequisite Toy Tasks [data] Jason Weston
- Blog: AND does not mean OR: Using Formal Languages to Study Language Models' Representations
- Blog: The Geometry of Prompting: Unveiling Distinct Mechanisms of Task Adaptation in Language Models
- Blog: Intrinsic Subspace Evaluation of Word Embedding Representations
- Blog: Practical Computational Power of Linear Transformers and Their Recurrent and Self-Referential Extensions
- Blog: Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence
- Blog: A Survey to Recent Progress Towards Understanding In-Context Learning
- Blog: Sparse Interventions in Language Models with Differentiable Masking
- Blog: mOthello: When Do Cross-Lingual Representation Alignment and Cross-Lingual Transfer Emerge in Multilingual Models?
- Blog: Deriving Neural Scaling Laws from the statistics of natural language
- Blog: Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning Process
- Blog: lilGym: Natural Language Visual Reasoning with Reinforcement Learning
- Blog: On the Bias of Next-Token Predictors Toward Systematically Inefficient Reasoning: A Shortest-Path Case Study
- Blog: Adversarial Robustness of In-Context Learning in Transformers for Linear Regression
- Blog: Rule Extrapolation in Language Modeling: A Study of Compositional Generalization on OOD Prompts
- Blog: In-Context Learning through the Bayesian Prism
- Blog: Semantic Training Signals Promote Hierarchical Syntactic Generalization in Transformers
- Blog: Investigating Transformer-Guided Chaining for Interpretable Natural Logic Reasoning
- Blog: When can in-context learning generalize out of task distribution?
- Blog: Learning to Organize Knowledge with N-Gram Machines
- Blog: The mechanistic basis of data dependence and abrupt learning in an in-context classification task
- Blog: Learning from Synthetic Data Improves Multi-hop Reasoning
- Blog: Causal language modeling can elicit search and reasoning capabilities on logic puzzles
- Blog: Teaching Machines to Read and Comprehend
- Blog: A Unified Perspective on Task Retrieval and Learning in In-Context Learning on Markov Data
- Blog: JFLD: A Japanese Benchmark for Deductive Reasoning Based on Formal Logic
- Blog: Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
- Blog: RL Post-Training Builds Compositional Reasoning Strategies
- Blog: Memory or Reasoning? Explore How LLMs Compute Mixed Arithmetic Expressions
- Blog: Do LLMs Build World Representations? Probing Through the Lens of State Abstraction
History & citation
Current · Permanent link
Cite this direction
Latest version (a living page):
@misc{lacuna_synthetic_to_natural_generalization_gap_6e21b2f1,
title = {synthetic-to-natural generalization gap},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/synthetic-to-natural-generalization-gap/rdr_6e21b2f16eb64c11808974d2b891bf83},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}This exact version (version 1, frozen text):
@misc{lacuna_synthetic_to_natural_generalization_gap_6e21b2f1_v1,
title = {synthetic-to-natural generalization gap},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/synthetic-to-natural-generalization-gap/rdr_6e21b2f16eb64c11808974d2b891bf83/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}