Bayesian reduction of attention posterior on copy task predicts first-order phase transition for softmax attention and second-order followed by crossover for linear attention.
Emergent Abilities in Large Language Models: A Survey
12 Pith papers cite this work, alongside 3 external citations. Polarity classification is still indexing.
12
Pith papers citing it
3
external citations · Pith
abstract
Large Language Models (LLMs) are leading a new technological revolution as one of the most promising research streams toward artificial general intelligence. The scaling of these models, accomplished by increasing the number of parameters and the magnitude of the training datasets, has been linked to various so-called emergent abilities that were previously unobserved. These emergent abilities, ranging from advanced reasoning and in-context learning to coding and problem-solving, have sparked an intense scientific debate: Are they truly emergent, or do they simply depend on external factors, such as training dynamics, the type of problems, or the chosen metric? What underlying mechanism causes them? Despite their transformative potential, emergent abilities remain poorly understood, leading to misconceptions about their definition, nature, predictability, and implications. In this work, we shed light on emergent abilities by conducting a comprehensive review of the phenomenon, addressing both its scientific underpinnings and real-world consequences. We first critically analyze existing definitions, exposing inconsistencies in conceptualizing emergent abilities. We then explore the conditions under which these abilities appear, evaluating the role of scaling laws, task complexity, pre-training loss, quantization, and prompting strategies. Our review extends beyond traditional LLMs and includes Large Reasoning Models (LRMs), which leverage reinforcement learning and inference-time search to amplify reasoning and self-reflection. However, emergence is not inherently positive. As AI systems gain autonomous reasoning capabilities, they also develop harmful behaviors, including deception, manipulation, and reward hacking. We highlight growing concerns about safety and governance, emphasizing the need for better evaluation frameworks and regulatory oversight.
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
An in-vitro study with synthetic languages finds cross-lingual transfer depends more on tokenization preserving reusable substructure than on lexical similarity or balance, with transfer emerging in stages.
Transformers on impossible-language variants show gradual grammatical sensitivity loss but sharp long-sentence generation failures, supporting generative deficiency as a link to non-attestation.
Mid-reasoning shifts in reasoning models are rare symptoms of unstable inference that seldom improve accuracy and do not reflect intrinsic self-correction.
A new 1,248-instance paired benchmark shows 18 vision-language models drop substantially in accuracy when images contain misleading visual cues.
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
A framework encodes observed trajectories and HD maps into tokens for frozen LLMs to perform spatio-temporal reasoning and predict future vehicle paths with a linear decoder.
An LLM-driven agent-based model with multi-round dialogue reproduces non-linear social influence patterns in vaccination opinion dynamics, with memory increasing resistance and prompt diversity increasing adoption.
AI models for automated short answer scoring show substantial mid-range quality degradation in expert agreement that improves with greater task-specific adaptation.
Proposes a fault-tolerance architecture for AI safety by analogizing unreliable AI artifacts to Byzantine nodes and applying consensus mechanisms.
Advanced language representations shape LLMs' schemas to improve knowledge activation and problem-solving.
citing papers explorer
-
Phase Transitions in Attention: A Bayesian Theory of Copy Head Emergence
Bayesian reduction of attention posterior on copy task predicts first-order phase transition for softmax attention and second-order followed by crossover for linear attention.
-
An In-Vitro Study on Cross-Lingual Generalization in Language Models
An in-vitro study with synthetic languages finds cross-lingual transfer depends more on tokenization preserving reusable substructure than on lexical similarity or balance, with transfer emerging in stages.
-
When transformers learn "impossible" languages, what do they learn?
Transformers on impossible-language variants show gradual grammatical sensitivity loss but sharp long-sentence generation failures, supporting generative deficiency as a link to non-attestation.
-
The Illusion of Insight in Reasoning Models
Mid-reasoning shifts in reasoning models are rare symptoms of unstable inference that seldom improve accuracy and do not reflect intrinsic self-correction.
-
MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs
A new 1,248-instance paired benchmark shows 18 vision-language models drop substantially in accuracy when images contain misleading visual cues.
-
Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
-
Frozen LLMs as Map-Aware Spatio-Temporal Reasoners for Vehicle Trajectory Prediction
A framework encodes observed trajectories and HD maps into tokens for frozen LLMs to perform spatio-temporal reasoning and predict future vehicle paths with a linear decoder.
-
A Large Language Model-Driven Agent-Based Modeling Framework with Multi-Round Communication for Simulating Vaccine Opinion Dynamics
An LLM-driven agent-based model with multi-round dialogue reproduces non-linear social influence patterns in vaccination opinion dynamics, with memory increasing resistance and prompt diversity increasing adoption.
-
Quality-Conditioned Agreement in Automated Short Answer Scoring: Mid-Range Degradation and the Impact of Task-Specific Adaptation
AI models for automated short answer scoring show substantial mid-range quality degradation in expert agreement that improves with greater task-specific adaptation.
-
A Byzantine Fault Tolerance Approach towards AI Safety
Proposes a fault-tolerance architecture for AI safety by analogizing unreliable AI artifacts to Byzantine nodes and applying consensus mechanisms.
-
Shaping Schema via Language Representation as the Next Frontier for LLM Intelligence Expanding
Advanced language representations shape LLMs' schemas to improve knowledge activation and problem-solving.
- Market Design for AI: Beyond the Copyright Binary