SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
super hub Canonical reference
Toolformer: Language Models Can Teach Themselves to Use Tools
Canonical reference. 87% of citing Pith papers cite this work as background.
145
Pith papers citing it
Background
87% of classified citations
abstract
Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. This is done in a self-supervised way, requiring nothing more than a handful of demonstrations for each API. We incorporate a range of tools, including a calculator, a Q\&A system, two different search engines, a translation system, and a calendar. Toolformer achieves substantially improved zero-shot performance across a variety of downstream tasks, often competitive with much larger models, without sacrificing its core language modeling abilities.
hub tools
citation-role summary
background 44
baseline 1
dataset 1
method 1
citation-polarity summary
claims ledger
- abstract Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. Thi
authors
co-cited works
representative citing papers
ShadowMerge exploits relation-channel conflicts to poison graph-based agent memory, achieving 93.8% average attack success rate on Mem0 and real-world datasets while bypassing existing defenses.
A language-model-driven agentic AI system autonomously executes multi-stage physics experiments at a production synchrotron light source, reducing preparation time by two orders of magnitude while upholding safety constraints.
RepairAgent autonomously repairs 164 bugs on Defects4J including 39 not fixed by prior techniques by treating an LLM as an agent that invokes tools via a finite state machine and dynamic prompts.
DSPy compiles short declarative programs into LM pipelines that self-optimize and outperform both standard few-shot prompting and expert-written chains on math, retrieval, and QA tasks.
Mind2Web is the first large-scale dataset of real-world web tasks for developing generalist language-guided agents that complete complex actions on diverse websites.
API-Bank is a new benchmark and training dataset for tool-augmented LLMs that shows fine-tuned models can approach GPT-3.5 tool-use effectiveness.
Sandboxed coding agents with text+image access match or outperform native omnimodal models on audio-video benchmarks by converting tasks into code-driven retrieval and processing.
MemGym unifies agent gyms into a memory benchmark with isolated scoring across tool-use, research, coding, and computer-use regimes plus a lightweight reward model for tractable coding evaluation.
Mobius Injection exploits semantic closure in LLM agents to enable single-message AbO-DDoS attacks achieving up to 51x call amplification and 229x latency inflation.
A new evaluation protocol shows agent memory reliability degrades variably with added irrelevant sessions depending on agent, memory interface, and scale.
MemFlow routes queries by intent to tiered memory operations, nearly doubling accuracy of a 1.7B SLM on long-horizon benchmarks compared to full-context baselines.
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
Developers use LLMs like ChatGPT mainly for knowledge acquisition and code generation at the detailed design level, reporting benefits such as better technology selection and early flaw detection alongside limitations like lengthy outputs, incorrect code, and hallucinations.
TraceToChain models LLM agent traces as absorbing DTMCs using automatic clustering and smoothed MLE, with KS and AIC validation, to reconcile pass@k, pass^k, and RDC as projections of a single first-passage success-time distribution.
A new 7x4 taxonomy organizes agentic AI security threats by architectural layer and persistence timescale, revealing under-explored upper layers and missing defenses after surveying 116 papers.
Moltbook operates as two largely separate layers: a dominant transactional token economy using protocols like MBC-20 and a thinner discursive conversation layer with only 3.6% agent overlap.
A novel function hijacking attack achieves 70-100% success rates in forcing specific function calls across five LLMs on the BFCL benchmark and is robust to context semantics.
Compositional selective specificity (CSS) decomposes generated answers into claims and emits each at the most specific level supported by evidence, raising overcommitment-aware utility from 0.846 to 0.913 on LongFact while retaining 0.938 specificity.
AgileLog introduces forkable shared logs with cheap forking and isolation to support AI agents on data streams.
Transactional Attention uses semantic sponsorship from anchor patterns to retain dormant critical tokens in KV caches, achieving 100% credential retrieval at 16 tokens where all prior methods fail.
LLMs match or exceed state-of-the-art traditional methods for stabilizing numerical expressions in scientific software, succeeding on 97.9% of expressions where baselines fail to improve accuracy, but struggle with control flow and high-precision literals.
Introduces LLM-mediated computing as a paradigm of reflective conversation and co-disclosure where the computer emerges through human-LLM interaction.
Reasoning LLMs with minimal tools for tree construction and analysis induce decision trees that outperform CART, compete with ensembles on low-resource tabular data, and provide human-readable reasoning traces.
citing papers explorer
-
SimWorld Studio: Automatic Environment Generation with Evolving Coding Agent for Embodied Agent Learning
SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
-
ShadowMerge: A Novel Poisoning Attack on Graph-Based Agent Memory via Relation-Channel Conflicts
ShadowMerge exploits relation-channel conflicts to poison graph-based agent memory, achieving 93.8% average attack success rate on Mem0 and real-world datasets while bypassing existing defenses.
-
Agentic AI for Multi-Stage Physics Experiments at a Large-Scale User Facility Particle Accelerator
A language-model-driven agentic AI system autonomously executes multi-stage physics experiments at a production synchrotron light source, reducing preparation time by two orders of magnitude while upholding safety constraints.
-
RepairAgent: An Autonomous, LLM-Based Agent for Program Repair
RepairAgent autonomously repairs 164 bugs on Defects4J including 39 not fixed by prior techniques by treating an LLM as an agent that invokes tools via a finite state machine and dynamic prompts.
-
DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines
DSPy compiles short declarative programs into LM pipelines that self-optimize and outperform both standard few-shot prompting and expert-written chains on math, retrieval, and QA tasks.
-
Mind2Web: Towards a Generalist Agent for the Web
Mind2Web is the first large-scale dataset of real-world web tasks for developing generalist language-guided agents that complete complex actions on diverse websites.
-
API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMs
API-Bank is a new benchmark and training dataset for tool-augmented LLMs that shows fine-tuned models can approach GPT-3.5 tool-use effectiveness.
-
Sandboxed Coding Agents are Competitive Omni-modal Task Solvers
Sandboxed coding agents with text+image access match or outperform native omnimodal models on audio-video benchmarks by converting tasks into code-driven retrieval and processing.
-
MemGym: a Long-Horizon Memory Environment for LLM Agents
MemGym unifies agent gyms into a memory benchmark with isolated scoring across tool-use, research, coding, and computer-use regimes plus a lightweight reward model for tractable coding evaluation.
-
Can a Single Message Paralyze the AI Infrastructure? The Rise of AbO-DDoS Attacks through Targeted Mobius Injection
Mobius Injection exploits semantic closure in LLM agents to enable single-message AbO-DDoS attacks achieving up to 51x call amplification and 229x latency inflation.
-
When Stored Evidence Stops Being Usable: Scale-Conditioned Evaluation of Agent Memory
A new evaluation protocol shows agent memory reliability degrades variably with added irrelevant sessions depending on agent, memory interface, and scale.
-
MemFlow: Intent-Driven Memory Orchestration for Small Language Model Agents
MemFlow routes queries by intent to tiered memory operations, nearly doubling accuracy of a 1.7B SLM on long-horizon benchmarks compared to full-context baselines.
-
BIM Information Extraction Through LLM-based Adaptive Exploration
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
-
Using LLMs in Software Design: An Empirical Study of GitHub and A Practitioner Survey
Developers use LLMs like ChatGPT mainly for knowledge acquisition and code generation at the detailed design level, reporting benefits such as better technology selection and early flaw detection alongside limitations like lengthy outputs, incorrect code, and hallucinations.
-
Measuring the Unmeasurable: Markov Chain Reliability for LLM Agents
TraceToChain models LLM agent traces as absorbing DTMCs using automatic clustering and smoothed MLE, with KS and AIC validation, to reconcile pass@k, pass^k, and RDC as projections of a single first-passage success-time distribution.
-
A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Framework
A new 7x4 taxonomy organizes agentic AI security threats by architectural layer and persistence timescale, revealing under-explored upper layers and missing defenses after surveying 116 papers.
-
The Platform Is Mostly Not a Platform: Token Economies and Agent Discourse on Moltbook
Moltbook operates as two largely separate layers: a dominant transactional token economy using protocols like MBC-20 and a thinner discursive conversation layer with only 3.6% agent overlap.
-
Breaking MCP with Function Hijacking Attacks: Novel Threats for Function Calling and Agentic Models
A novel function hijacking attack achieves 70-100% success rates in forcing specific function calls across five LLMs on the BFCL benchmark and is robust to context semantics.
-
Answer Only as Precisely as Justified: Calibrated Claim-Level Specificity Control for Agentic Systems
Compositional selective specificity (CSS) decomposes generated answers into claims and emits each at the most specific level supported by evidence, raising overcommitment-aware utility from 0.846 to 0.913 on LongFact while retaining 0.938 specificity.
-
AgileLog: A Forkable Shared Log for Agents on Data Streams
AgileLog introduces forkable shared logs with cheap forking and isolation to support AI agents on data streams.
-
Transactional Attention: Semantic Sponsorship for KV-Cache Retention
Transactional Attention uses semantic sponsorship from anchor patterns to retain dormant critical tokens in KV caches, achieving 100% credential retrieval at 16 tokens where all prior methods fail.
-
Assessing Large Language Models for Stabilizing Numerical Expressions in Scientific Software
LLMs match or exceed state-of-the-art traditional methods for stabilizing numerical expressions in scientific software, succeeding on 97.9% of expressions where baselines fail to improve accuracy, but struggle with control flow and high-precision literals.
-
Co-Disclosing the Computer: LLM-Mediated Computing through Reflective Conversation
Introduces LLM-mediated computing as a paradigm of reflective conversation and co-disclosure where the computer emerges through human-LLM interaction.
-
Talking Trees: Reasoning-Assisted Induction of Decision Trees for Tabular Data
Reasoning LLMs with minimal tools for tree construction and analysis induce decision trees that outperform CART, compete with ensembles on low-resource tabular data, and provide human-readable reasoning traces.
-
$\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
τ-bench shows state-of-the-art agents like GPT-4o succeed on under 50% of tool-using, rule-following tasks and are inconsistent across repeated trials.
-
LLM Agents can Autonomously Exploit One-day Vulnerabilities
GPT-4 LLM agents autonomously exploit 87% of tested one-day vulnerabilities when given CVE descriptions, far outperforming other models and tools.
-
GAIA: a benchmark for General AI Assistants
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
-
Large Language Models as Optimizers
Large language models can optimize by being prompted with histories of past solutions and scores to propose better ones, producing prompts that raise accuracy up to 8% on GSM8K and 50% on Big-Bench Hard over human-designed baselines.
-
SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Models
SciBench shows current LLMs reach at most 43.22% accuracy on curated collegiate scientific problems and reveals no prompting strategy dominates across all required skills.
-
LLM+P: Empowering Large Language Models with Optimal Planning Proficiency
LLM+P lets LLMs solve planning problems optimally by converting them to PDDL for classical planners and back to natural language.
-
Reflexion: Language Agents with Verbal Reinforcement Learning
Reflexion lets LLM agents improve via stored verbal reflections on task feedback, reaching 91% pass@1 on HumanEval and outperforming prior GPT-4 results.
-
ViperGPT: Visual Inference via Python Execution for Reasoning
ViperGPT generates executable Python code to compose pre-trained vision-and-language modules into programs that answer visual queries, reaching state-of-the-art results with no additional training.
-
Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks
PoT prompting improves numerical reasoning by having language models write programs executed by a computer instead of performing calculations in natural language chains of thought, with an average 12% gain over CoT.
-
The Self-Correction Illusion: LLMs Correct Others but Not Themselves
Relabeling an identical erroneous claim from the model's own thought role to an external chat role increases explicit correction rates by 23-93 percentage points across 13 model-domain cells, indicating a chat-template artifact rather than a cognitive deficit.
-
EVE-Agent: Evidence-Verifiable Self-Evolving Agents
EVE-Agent adds an evidence verifier to the proposer-solver loop that rewards spans by marginal accuracy gain, producing self-generated but inspectable training examples for search agents.
-
PULSE: Agentic Investigation with Passive Sensing for Proactive Intervention in Cancer Survivorship
PULSE demonstrates that agentic LLM-based investigation of passive smartphone sensing data achieves balanced accuracies of 0.743 (with diary) and 0.713 (sensing-only) for predicting emotion regulation desire and intervention availability in 50 cancer survivors.
-
RooAgent: An LLM Agent for Root-Based High Energy Physics Analysis
RooAgent provides an LLM agent interface that translates natural-language prompts into calls to PyROOT analysis functions for high energy physics tasks, with support for multiple AI backends and tested on ZH simulations and ATLAS open data.
-
Resolving Action Bottleneck: Agentic Reinforcement Learning Informed by Token-Level Energy
ActFocus resolves the action bottleneck in agentic RL by reweighting token gradients toward action tokens using observed reward variance and an energy-based uncertainty term, outperforming PPO and GRPO by up to 65 percentage points.
-
LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scaling
A metacognitive harness uses LLMs' pre- and post-solution self-monitoring signals to control test-time reasoning, raising pooled accuracy from 48.3% to 56.9% on text, code, and multimodal benchmarks.
-
Towards Long-horizon Embodied Agents with Tool-Aligned Vision-Language-Action Models
VLAs-as-Tools pairs a VLM planner with specialized VLA executors via a new interface and Tool-Aligned Post-Training to raise long-horizon robot success rates on LIBERO-Long and RoboTwin benchmarks.
-
Domain Restriction via Multi SAE Layer Transitions
Multi-layer SAE transitions capture domain-specific signatures that distinguish OOD texts in Gemma-2 models.
-
From Controlled to the Wild: Evaluation of Pentesting Agents for the Real-World
A practical evaluation protocol for AI pentesting agents that uses validated vulnerability discovery, LLM semantic matching, and bipartite scoring to assess performance in realistic, complex targets.
-
Deterministic vs. LLM-Controlled Orchestration for COBOL-to-Python Modernization
Deterministic orchestration matches LLM-controlled methods in COBOL-to-Python translation accuracy but improves worst-case robustness, reduces run-to-run variability, and cuts token consumption by up to 3.5 times.
-
SkillGen: Verified Inference-Time Agent Skill Synthesis
SkillGen synthesizes auditable skills from agent trajectories via contrastive induction on successes and failures, then verifies net performance impact by comparing outcomes with and without the skill on identical tasks.
-
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.
-
SkillLens: Adaptive Multi-Granularity Skill Reuse for Cost-Efficient LLM Agents
SkillLens organizes skills into policies-strategies-procedures-primitives layers, retrieves via degree-corrected random walk, and uses a verifier for local adaptation, yielding up to 6.31 pp gains on MuLocbench and raising ALFWorld success from 45% to 51.31%.
-
Unsafe by Flow: Uncovering Bidirectional Data-Flow Risks in MCP Ecosystem
MCP-BiFlow detects 93.8% of known bidirectional data-flow vulnerabilities in MCP servers and identifies 118 confirmed issues across 87 real-world servers from a scan of 15,452 repositories.
-
PaT: Planning-after-Trial for Efficient Test-Time Code Generation
PaT defers planning until after failed trials in LLM code generation, enabling heterogeneous cheap-plus-powerful model setups that match large-model performance at roughly 69% lower cost.
-
BioTool: A Comprehensive Tool-Calling Dataset for Enhancing Biomedical Capabilities of Large Language Models
BioTool dataset enables fine-tuning a 4B-parameter LLM to outperform GPT-5.1 in biomedical tool calling while improving downstream answer quality per human experts.
-
AgentWard: A Lifecycle Security Architecture for Autonomous AI Agents
AgentWard organizes stage-specific security controls with cross-layer coordination to intercept threats across the full lifecycle of autonomous AI agents.