Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Canaverde, Beatriz, Alves, Duarte M., Pombal, José, Attanasio, Giuseppe, Martins, André F. T.
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.06353
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914542287585280
author Canaverde, Beatriz
Alves, Duarte M.
Pombal, José
Attanasio, Giuseppe
Martins, André F. T.
author_facet Canaverde, Beatriz
Alves, Duarte M.
Pombal, José
Attanasio, Giuseppe
Martins, André F. T.
contents In a conversation, a helpful assistant must reliably follow user directives, even as they refine, modify, or contradict earlier requests. Yet most instruction-following benchmarks focus on single-turn or short multi-turn scenarios, leaving open how well models handle long-horizon instruction-following tasks. To bridge this gap, we present SEQUOR, an automatic benchmark for evaluating constraint adherence in long multi-turn conversations. SEQUOR consists of simulated persona-driven interactions built with constraints extracted from real-world conversations. Our results show that even when following a single constraint, instruction-following accuracy consistently decreases as the conversation grows longer, with drops exceeding 11%. This decline becomes larger when models have to follow multiple constraints simultaneously, reducing their accuracy by over 40%. In scenarios where constraints are added or replaced at arbitrary points of the conversation, model accuracy decreases by more than 9%. Taken together, our results reveal that current models still struggle to follow user instructions in multi-turn conversations, and provide a way for better measuring instruction-following capabilities in assistants.
format Preprint
id arxiv_https___arxiv_org_abs_2605_06353
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SEQUOR: A Multi-Turn Benchmark for Realistic Constraint Following
Canaverde, Beatriz
Alves, Duarte M.
Pombal, José
Attanasio, Giuseppe
Martins, André F. T.
Computation and Language
In a conversation, a helpful assistant must reliably follow user directives, even as they refine, modify, or contradict earlier requests. Yet most instruction-following benchmarks focus on single-turn or short multi-turn scenarios, leaving open how well models handle long-horizon instruction-following tasks. To bridge this gap, we present SEQUOR, an automatic benchmark for evaluating constraint adherence in long multi-turn conversations. SEQUOR consists of simulated persona-driven interactions built with constraints extracted from real-world conversations. Our results show that even when following a single constraint, instruction-following accuracy consistently decreases as the conversation grows longer, with drops exceeding 11%. This decline becomes larger when models have to follow multiple constraints simultaneously, reducing their accuracy by over 40%. In scenarios where constraints are added or replaced at arbitrary points of the conversation, model accuracy decreases by more than 9%. Taken together, our results reveal that current models still struggle to follow user instructions in multi-turn conversations, and provide a way for better measuring instruction-following capabilities in assistants.
title SEQUOR: A Multi-Turn Benchmark for Realistic Constraint Following
topic Computation and Language
url https://arxiv.org/abs/2605.06353