arXiv subjects

Devansh Singh

Publications and source records attributed to Devansh Singh.

4 recordsLinked to original sources

arXiv2026

Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverse

Agentic AI systems are increasingly adopting automated pipelines that integrate multiple tools. While prior research and benchmarks have studied about task success and task completion of these agentic systems, the research about agent to tool interaction, specifically in biology agentic workflow is limited. This study investigates specific failures in agent to tool interaction where a tool invocation appears successful, some or all of the information or functionality from the tool via API/ wrapper is incomplete or missing and there are no communications / notifications to the user or the agent about such missing information. We call this a silent failures as the user or the agents are not aware that such failure has occurred. For the purposes of this study we developed an audit mechanism to identify such silent failures in Agent to tool interaction, by examining 15 scientific tools (and their associated API documentation and tool documentations) integrated within ToolUniverse environment (ToolUniverse serves as our experimental environment rather than the object of the study itself). We structure our study around 7 failure locus characterising where the failure occurs in the chain. We observed 91 failures (manually validated post LLM based candidate discovery and automated testing), most frequent of them being missing data or fields and inconsistencies in search, filtering or ranking criteria. Most of the 91 failures occurred in API layer (51) or wrapper layer (25), with a potential of silent failure amplification downstream. The results show that silent failures originate upstream of the event and propagate downstream into apparently valid scientific outputs. We propose a concept of contextual reliability to handle such failures and suggest mechanisms for testing, disclosing, monitoring, and measuring such failures across the agent-tool interaction pipeline.

cs.AI↗

arXiv2026

Beyond Accuracy and Surface Fluency: Risk-Sensitive Evaluation of LLMs for Legal Clause Generation

Large language models (LLMs) are increasingly used to draft contractual language, yet conventional accuracy or preference-based evaluations are poorly matched to legal drafting. A clause may be fluent and stylistically polished while still omitting an essential carve-out, allocating risk in an unenforceable way, assuming an inapplicable jurisdiction, or exposing a party to regulatory liability. This paper presents a empirical study design and framework for evaluating LLM-generated contract clauses. The study evaluates four models - Claude Haiku 4.5, Gemini 2.5 Flash Lite, GPT 5.4 Nano, and Qwen 3.5 Flash, across 22 contract clause categories and 34 legally-motivated failure modes. We combine two evaluation frameworks: CLAUSE, which classifies prompts by legal function and failure target, and LENS-CRAFT, which scores outputs across nine legal-quality dimensions. Instead of averaging dimension scores, the study applies a Max Severity Principle so that a single legally decisive defect remains visible. The paper provides the evaluation protocol, taxonomy, analysis plan, and a results structure for reporting empirical findings. We argue that legal AI evaluation should move beyond aggregate accuracy toward clause-specific, failure-mode-driven, and risk-sensitive assessment.

cs.CL↗

arXiv2025

Unmasking the Reality of PII Masking Models: Performance Gaps and the Call for Accountability

Privacy Masking is a critical concept under data privacy involving anonymization and de-anonymization of personally identifiable information (PII). Privacy masking techniques rely on Named Entity Recognition (NER) approaches under NLP support in identifying and classifying named entities in each text. NER approaches, however, have several limitations including (a) content sensitivity including ambiguous, polysemic, context dependent or domain specific content, (b) phrasing variabilities including nicknames and alias, informal expressions, alternative representations, emerging expressions, evolving naming conventions and (c) formats or syntax variations, typos, misspellings. However, there are a couple of PII datasets that have been widely used by researchers and the open-source community to train models on PII detection or masking. These datasets have been used to train models including Piiranha and Starpii, which have been downloaded over 300k and 580k times on HuggingFace. We examine the quality of the PII masking by these models given the limitations of the datasets and of the NER approaches. We curate a dataset of 17K unique, semi-synthetic sentences containing 16 types of PII by compiling information from across multiple jurisdictions including India, U.K and U.S. We generate sentences (using language models) containing these PII at five different NER detection feature dimensions - (1) Basic Entity Recognition, (2) Contextual Entity Disambiguation, (3) NER in Noisy & Real-World Data, (4) Evolving & Novel Entities Detection and (5) Cross-Lingual or multi-lingual NER) and 1 in adversarial context. We present the results and exhibit the privacy exposure caused by such model use (considering the extent of lifetime downloads of these models). We conclude by highlighting the gaps in measuring performance of the models and the need for contextual disclosure in model cards for such models.

cs.CL↗

arXiv2021

Rational AI: A comparison of human and AI responses to triggers of economic irrationality in poker

Humans exhibit irrational decision-making patterns in response to environmental triggers, such as experiencing an economic loss or gain. In this paper we investigate whether algorithms exhibit the same behavior by examining the observed decisions and latent risk and rationality parameters estimated by a random utility model with constant relative risk-aversion utility function. We use a dataset consisting of 10,000 hands of poker played by Pluribus, the first algorithm in the world to beat professional human players and find (1) Pluribus does shift its playing style in response to economic losses and gains, ceteris paribus; (2) Pluribus becomes more risk-averse and rational following a trigger but the humans become more risk-seeking and irrational; (3) the difference in playing styles between Pluribus and the humans on the dimensions of risk-aversion and rationality are particularly differentiable when both have experienced a trigger. This provides support that decision-making patterns could be used as "behavioral signatures" to identify human versus algorithmic decision-makers in unlabeled contexts.

econ.TH↗