Pith. sign in

REVIEW 6 cited by

Do LLMs Know When to NOT Answer? Investigating Abstention Abilities of Large Language Models

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 2407.16221 v2 pith:LJNLCYJG submitted 2024-07-23 cs.CL

Do LLMs Know When to NOT Answer? Investigating Abstention Abilities of Large Language Models

classification cs.CL
keywords abstentionmodelspromptinganswerapproachblack-boxcapabilityconfusion
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Abstention Ability (AA) is a critical aspect of Large Language Model (LLM) reliability, referring to an LLM's capability to withhold responses when uncertain or lacking a definitive answer, without compromising performance. Although previous studies have attempted to improve AA, they lack a standardised evaluation method and remain unsuitable for black-box models where token prediction probabilities are inaccessible. This makes comparative analysis challenging, especially for state-of-the-art closed-source commercial LLMs. This paper bridges this gap by introducing a black-box evaluation approach and a new dataset, Abstain-QA, crafted to rigorously assess AA across varied question types (answerable and unanswerable), domains (well-represented and under-represented), and task types (fact centric and reasoning). We also propose a new confusion matrix, the ''Answerable-Unanswerable Confusion Matrix'' (AUCM) which serves as the basis for evaluating AA, by offering a structured and precise approach for assessment. Finally, we explore the impact of three prompting strategies-Strict Prompting, Verbal Confidence Thresholding, and Chain-of-Thought (CoT)-on improving AA. Our results indicate that even powerful models like GPT-4, Mixtral 8x22b encounter difficulties with abstention; however, strategic approaches such as Strict prompting and CoT can enhance this capability.

discussion (0)

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

Forward citations

Cited by 6 Pith papers

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

  1. Partial Evidence Bench: Benchmarking Authorization-Limited Evidence in Agentic Systems

    cs.AI 2026-05 unverdicted novelty 7.0

    Partial Evidence Bench is a deterministic benchmark that measures agent correctness, completeness awareness, gap-report quality, and unsafe overclaiming in authorization-constrained evidence environments.

  2. Don't Start What You Can't Finish: A Counterfactual Audit of Support-State Triage in LLM Agents

    cs.AI 2026-04 unverdicted novelty 7.0

    LLM agents overcommit on non-complete tasks at 41.7% unless given explicit support-state categories, which raise typed deferral accuracy to 91.7%.

  3. BAS: A Decision-Theoretic Approach to Evaluating Large Language Model Confidence

    cs.CL 2026-04 unverdicted novelty 7.0

    BAS aggregates utility from an answer-or-abstain model across risk thresholds and is uniquely maximized by truthful confidence estimates.

  4. Causal Evidence that Language Models use Confidence to Drive Behavior

    cs.LG 2026-03 unverdicted novelty 6.0

    Language models deploy multidimensional internal confidence representations and threshold-based policies to control abstention behavior, with causal support from activation steering experiments.

  5. Auto-DSM Under the Lens: A Black-Box Evaluation Framework for LLM-Based DSM Generation

    cs.AI 2026-07 conditional novelty 5.0

    A black-box evaluation framework with multi-run metrics reveals that LLM-based DSM generation is sensitive to phrasing, exhibits hallucination, and fails to consistently acknowledge knowledge boundaries.

  6. When Retrieval Metrics Mislead: Measuring Policy Signal in Long-Horizon Tool-Use Agents

    cs.CL 2026-06 unverdicted novelty 4.0

    Retrieved clauses yield macro-F1 within 0.02 of gold clauses (wide CI) despite 7% rank-1 exact match, indicating exact-match recall underestimates policy utility in this agent benchmark.