REVIEW 8 cited by
The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore Non-Determinism
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
The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore Non-Determinism
read the original abstract
Current evaluations of large language models (LLMs) often overlook non-determinism, typically focusing on a single output per example. This limits our understanding of LLM performance variability in real-world applications. Our study addresses this issue by exploring key questions about the performance differences between greedy decoding and sampling, identifying benchmarks' consistency regarding non-determinism, and examining unique model behaviors. Through extensive experiments, we observe that greedy decoding generally outperforms sampling methods for most evaluated tasks. We also observe consistent performance across different LLM sizes and alignment methods, noting that alignment can reduce sampling variance. Moreover, our best-of-N sampling approach demonstrates that smaller LLMs can match or surpass larger models such as GPT-4-Turbo, highlighting the untapped potential of smaller LLMs. This research shows the importance of considering non-determinism in LLM evaluations and provides insights for future LLM development and evaluation.
Forward citations
Cited by 8 Pith papers
-
Verified Detection and Prevention of Concurrency Anomalies in Multi-Agent Large Language Model Systems
Formalizes four concurrency anomalies in multi-agent LLM systems and mechanically verifies a hierarchy of sound detectors and preventions realized in Rust runtimes using TLA+ and Verus.
-
Verifier-free Test-Time Sampling for Vision-Language-Action Models
A verifier-free test-time sampling method for vision-language-action models that selects actions by KL divergence to a condition-masked reference distribution, improving task success rates.
-
The One-Word Census: Answer-Choice Conformity Across 44 Language Models
Forty-four language models asked to name one thing per category converge on the same modal answers far more than people do, with newest flagships most conformist and persona-tuned models most divergent.
-
The One-Word Census: Answer-Choice Conformity Across 44 Language Models
Across 31 open one-word categories, 44 LMs converge extremely (often >80% on one answer), with newest flagships most conformist and persona-tuned models most divergent.
-
Introducing Background Temperature to Characterise Hidden Randomness in Large Language Models
LLMs show implementation-induced randomness even at T=0 that can be characterized as an effective background temperature T_bg estimated via an ideal reference system.
-
LLM-based Generation of Semantically Diverse and Realistic Domain Model Instances
LLMs with two prompting strategies and model validation tools produce mostly syntactically correct, conforming, semantically realistic and diverse instances of UML class diagrams.
-
Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
Repeated sampling scales problem coverage log-linearly with sample count, improving SWE-bench Lite performance from 15.9% to 56% using 250 samples.
-
Cross-Model Consistency of AI-Generated Exercise Prescriptions: A Repeated Generation Study Across Three Large Language Models
Three LLMs exhibit distinct consistency profiles in repeated exercise prescription generation, with GPT-4.1 producing unique but semantically stable outputs while Gemini 2.5 Flash achieves high similarity through text...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.