Pith. sign in

REVIEW 1 cited by

Are LLMs All You Need for Task-Oriented Dialogue?

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2304.06556 v2 pith:ZB7GIEBS submitted 2023-04-13 cs.CL

Are LLMs All You Need for Task-Oriented Dialogue?

classification cs.CL
keywords abilitydialoguellmsbeliefinteractmodelsstatetask-oriented
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Instructions-tuned Large Language Models (LLMs) gained recently huge popularity thanks to their ability to interact with users through conversation. In this work we aim to evaluate their ability to complete multi-turn tasks and interact with external databases in the context of established task-oriented dialogue benchmarks. We show that for explicit belief state tracking, LLMs underperform compared to specialized task-specific models. Nevertheless, they show ability to guide the dialogue to successful ending if given correct slot values. Furthermore this ability improves with access to true belief state distribution or in-domain examples.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Building Task Bots with Self-learning for Enhanced Adaptability, Extensibility, and Factuality

    cs.CL 2025-08 conditional novelty 2.0

    A thesis that combines self-learning from dialog logs, schema-guided prompting, and self-aligned factuality to build task bots with minimal human intervention.