Semantic Early-Stopping for Iterative LLM Agent Loops

• URL: http://arxiv.org/abs/2606.27009v1
• Date: Thu, 25 Jun 2026 13:24:21 GMT
• Title: Semantic Early-Stopping for Iterative LLM Agent Loops
• Abstract summary: We study semantic early-stopping: the loop halts when consecutive draft embeddings stop changing in meaning.<n>We generate each question's full trajectory once, replay every stopping policy over the identical drafts, and cache every LLM-judge call.<n>An oracle that selects the best round attains +0.115 Information Score over every practical policy.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multi-agent large language model (LLM) loops, for example a Writer that drafts and a Critic that revises, are almost always terminated by a fixed iteration cap (max_iterations). This is a syntactic kill-switch: it is blind to whether the answer is still improving, so it over-spends tokens on easy inputs and truncates hard ones. We study semantic early-stopping: the loop halts when consecutive draft embeddings stop changing in meaning (cosine distance with a patience window) and the answer's measured quality stops improving. Our work makes three contributions. First, an honest theoretical footing: we prove deterministic termination and well-definedness and machine-check these claims, while treating the convergence of the distance sequence as an empirically tested conjecture rather than a (previously over-claimed) Banach contraction. Second, a judge-efficient evaluation protocol: we generate each question's full trajectory once, replay every stopping policy over the identical drafts, and cache every LLM-judge call, yielding a strictly paired efficiency-versus-quality comparison at low cost; we further separate operational tokens (charged to a policy) from evaluation tokens (a measurement instrument). Third, an empirical study on multi-hop retrieval-augmented question answering (HotpotQA). On the 60-question test split, a judge-free semantic stopper reduces operational tokens by 38% relative to max_iterations at parity quality (Delta-IS = -0.004, p = 0.81), whereas the full quality-gated variant is counter-productive because its per-round judging dominates cost. An oracle that selects the best round attains +0.115 Information Score over every practical policy (p ~ 4e-11), reframing the problem from "when to stop" (easy) to "which round is best" (open).

Related papers

• Partial Contracts Suffice: Sound, LLM-Inferred Regression Verification [7.004725450780227]
  We present the first contract-based regression verification tool.<n>Contract soundness is ensured by proving all function versions match the behavior.<n>We ask whether a partial, caller-sufficient contract, rather than a full behavioral specification, is enough.
  arXiv  Detail & Related papers  (2026-07-11T12:40:22Z)
• Re-feeding Is Not Replaying: Measuring Replay Noise in Counterfactual Token-Credit Estimation [0.0]
  Per-token counterfactual credit estimation asks which token in a language-model rollout caused the final answer to be right or wrong.<n>We measure what that assumption costs on a stock inference engine, with a three-pass design.<n>Across six configurations and three models, re-feeding changes the credit estimate at rates 14-28 percentage points above the replica floor.
  arXiv  Detail & Related papers  (2026-06-14T06:09:16Z)
• The Last Word Often Wins: A Format Confound in Chain-of-Thought Corruption Studies [0.0]
  Corruption studies infer which steps are computationally important'' from accuracy loss when steps are corrupted.<n>We show that when benchmark chains end with an explicit terminal answer line, these tests largely measure emphanswer placement rather than where intermediate computation is carried out.
  arXiv  Detail & Related papers  (2026-05-11T16:26:50Z)
• Two Calls, Two Moments, and the Vote-Accuracy Curve of Repeated LLM Inference [4.28787537081191]
  We study the binary correctness layer of repeated LLM inference under conditional-i.i.d. calls.<n>Every fixed majority-vote budget has a sharp distribution-free two-call interval.<n>Experiments on LLM calls over QNLI and QQP show that empirical three- and five-vote accuracies are contained in the projected two-call regions.
  arXiv  Detail & Related papers  (2026-05-05T05:40:09Z)
• Efficient Test-Time Inference via Deterministic Exploration of Truncated Decoding Trees [68.04613115686509]
  Self-consistency boosts inference-time performance by sampling multiple reasoning traces in parallel and voting.<n>We propose Distinct Leafion (DLE), a deterministic decoding method that treats truncated sampling as a pruned tree and systematically enumerates distinct leaves instead of sampling with replacement.<n>DLE explores higher-quality reasoning traces, yielding better performance on math, coding, and general reasoning tasks.
  arXiv  Detail & Related papers  (2026-04-22T12:42:03Z)
• Test-Time Scaling with Diffusion Language Models via Reward-Guided Stitching [66.39914384073145]
  We propose a self-consistency framework that turns cheap diffusion-sampled reasoning into a reusable pool of step-level candidates.<n>We find that step-level recombination is most beneficial on harder problems.<n>Our training-free framework improves average accuracy by up to 2 across six math and coding tasks.
  arXiv  Detail & Related papers  (2026-02-26T11:08:39Z)
• Global Resolution: Optimal Multi-Draft Speculative Sampling via Convex Minimization [1.2674961594128336]
  We devise an algorithm for optimal speculative sampling when the $n$ tokens are chosen from a single draft model.<n>Our findings give the first multi-draft algorithm with 90% acceptance and under 100 ms of overhead per generated token with negligible deviation from the target model distribution.
  arXiv  Detail & Related papers  (2025-11-19T21:59:43Z)
• Rethinking Thinking Tokens: LLMs as Improvement Operators [80.12087211785949]
  Reasoning training incentivizes LLMs to produce long chains of thought (long CoT), which allows them to explore solution strategies with self-checking.<n>This results in higher accuracy, but inflates context length, token/compute cost, and answer latency.<n>We ask: Can current models leverage their metacognition to provide other combinations on this Pareto frontier?<n>We identify an interesting inference family Parallel-Distill-Refine (PDR), which performs the following: (i) generate diverse drafts in parallel; (ii) distill them into a bounded, textual workspace; and (iii) refine conditioned on this workspace
  arXiv  Detail & Related papers  (2025-10-01T17:08:59Z)
• Fractured Chain-of-Thought Reasoning [61.647243580650446]
  We introduce Fractured Sampling, a unified inference-time strategy that interpolates between full CoT and solution-only sampling.<n>We show that Fractured Sampling consistently achieves superior accuracy-cost trade-offs, yielding steep log-linear scaling gains in Pass@k versus token budget.
  arXiv  Detail & Related papers  (2025-05-19T11:30:41Z)
• Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling [90.86991492288487]
  evaluating constraint on every token can be prohibitively expensive.<n> LCD can distort the global distribution over strings, sampling tokens based only on local information.<n>We show that our approach is superior to state-of-the-art baselines.
  arXiv  Detail & Related papers  (2025-04-07T18:30:18Z)
• Prune Spatio-temporal Tokens by Semantic-aware Temporal Accumulation [89.88214896713846]
  STA score considers two critical factors: temporal redundancy and semantic importance. We apply the STA module to off-the-shelf video Transformers and Videowins. Results: Kinetics-400 and Something-Something V2 achieve 30% overshelf reduction with a negligible 0.2% accuracy drop.
  arXiv  Detail & Related papers  (2023-08-08T19:38:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.