ArtificialIntelligence.io

The Signal

Everything that matters in AI, with our take.

Updated through the day. Every headline links straight to the source. The two lines underneath are ours.

arXiv cs.CLPaper

If It's Not Buggy, Don't Fix It: On the Dynamics of Iterative Bug-fixing with LLMs

This has real implications for autonomous code agents in production. The mechanistic finding about a steering vector for editing propensity suggests the problem is tractable, not fundamental. If you're deploying LLM-based code repair or CI/CD automation, you need guardrails: verify before committing, cap iteration depth, and monitor for the pseudo-cycle pattern. This moves LLM code tools from "nice to have" to "needs gates".

arXiv cs.AIPaper

CodeTD: Topology of Attention Detects Hallucinations in Code LLMs

Pre-execution hallucination detection is valuable because code you can't verify is code you can't ship. Topological analysis of attention is a novel lens that generalizes across models and languages. For teams using code LLMs in production, this is worth testing: it runs without execution and could save you from deploying broken code. The transferability across benchmarks is a real strength.

arXiv cs.CLPaper

Measuring LLM Sycophancy under Sustained Multi-Turn Pressure

This closes a real evaluation gap. Short-horizon sycophancy tests miss the failure mode that matters in real customer service, support, and domain expert use cases. All four production systems tested deteriorate under sustained pressure. If you're building systems where the model's reliability on corrections is safety-critical, you need to know that current models aren't ready for that without guardrails. The reasoning trace analysis hints at a fix: the right answer is there, the model just chooses to abandon it.

arXiv cs.LGPaper

How to Speculate about Uncertainty in Agentic Coding? A Draft-Model Gate Method

The problem is real: LLM coding agents fail confidently and expensive retries happen post-execution. This method extracts failure likelihood from output text alone, which is practical since you can't always access internal model state, especially with closed-source agents. The pre-execution veto gate is a concrete deployment pattern. For teams running large-scale agentic coding, this is worth testing against your own error patterns.

arXiv cs.CLPaper

LLM Judges as Raters: A Pre-Registered Audit of Severity, Halo, Reliability, and Version Instability in LLM Essay Scoring on Public Corpora

This is the warning you needed before deploying LLMs as graders at scale. The rater-effects methodology is borrowed from educational measurement but applied rigorously to LLMs, and the findings are damning: LLMs drift between versions, show bias patterns like halo, and disagree with each other more than humans disagree. If you're using LLMs to score anything high-stakes, this paper just became your liability shield. Read the method; it's reproducible.

arXiv cs.AIPaper

Clean Engineering, Unstable Measurement: A Preregistered Reliability Failure of Black-Box LLM Observers on Shared Endpoints

This is a systems problem that invalidates a lot of recent benchmarking work. If your leaderboards or training pipelines rely on LLM judges, your measurements are noisier than you think. The paper identifies three specific failure modes: label-to-meaning drift, sub-noise-floor variance, and non-determinism on identical inputs. Audit your evaluation pipeline now.

arXiv cs.AIPaper

Epistemic Warrant for LLM Recommendations: Characterizing the Basis for Reliance When Ground Truth Is Unavailable

The problem is real: when ground truth isn't available, how do you know if a model's answer is trustworthy? The answer is clever but limited. A four-tier reliance certificate helps, but it still requires humans to validate. This is more useful as a conceptual framework for your own product than as deployable infrastructure. The gap between theoretically sound and operationally useful remains.

arXiv cs.AIPaper

Diagnosing with Insights: Structured Analysis of Agent Failures via Behavioral Abstractions

Agent debugging is a real friction point right now, and this addresses it with a concrete system. The approach of using structured abstractions plus neural invariants is cleaner than pure LLM-as-judge diagnosis. If you're shipping agents in production and drowning in opaque failure logs, this method deserves a test run. It won't solve everything, but it's directional.

arXiv cs.CLPaper

Dutch Books for Language Models

Language models are widely used for probabilistic judgment (market forecasts, risk estimates, event likelihoods) and most users assume the probabilities are internally consistent. They're not. The incoherence grows under certain conditions. If you're using LLM outputs for quantitative decision-making, you need to know this and either add calibration or treat the numbers as rough signals only.

arXiv cs.LGPaper

On the Resilience of Text-to-Video Diffusion Models to Hardware Faults

If you're running T2V models in production at scale, this matters. Memory faults are worse than compute faults, bfloat16 is riskier than alternatives, and the scary part is that some faults cause semantic changes, not just noise. This is the kind of systems reliability work that becomes critical as video generation moves from hobbyist to production. Test your deployment stack against these fault modes.

arXiv cs.CLPaper

Calibrated Enough to Know, Not Calibrated to Act: Fabricated Evidence Makes LLM Agents Commit to the Unknowable

This is a sharp finding for anyone deploying agents in financial, forecasting, or advisory contexts: the models aren't fooled by false information so much as by the appearance of authority. Stated confidence scores don't move even as behavior swings 48 points, meaning you can't rely on a model's self-reported uncertainty to catch this failure. Anyone building agents that consume dashboards or reports needs a guardrail that checks provenance, not just plausibility.

arXiv cs.CLPaper

When Text and Numbers Disagree: Evidence Arbitration in Large Language Models

This matters for anyone building agents that pull from mixed sources, financial dashboards, monitoring systems, tool outputs feeding a summarizer. The finding that models over-trust recent data and external forecasts even against explicit reliability signals is exactly the kind of failure mode that shows up quietly in production and causes bad decisions. If your pipeline reconciles numbers and text automatically, this is worth testing against your own models before you trust the arbitration.