REVIEW 18 cited by
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
Time Travel in LLMs: Tracing Data Contamination in Large Language Models
read the original abstract
Data contamination, i.e., the presence of test data from downstream tasks in the training data of large language models (LLMs), is a potential major issue in measuring LLMs' real effectiveness on other tasks. We propose a straightforward yet effective method for identifying data contamination within LLMs. At its core, our approach starts by identifying potential contamination at the instance level; using this information, our approach then assesses wider contamination at the partition level. To estimate contamination of individual instances, we employ "guided instruction:" a prompt consisting of the dataset name, partition type, and the random-length initial segment of a reference instance, asking the LLM to complete it. An instance is flagged as contaminated if the LLM's output either exactly or nearly matches the latter segment of the reference. To understand if an entire partition is contaminated, we propose two ideas. The first idea marks a dataset partition as contaminated if the average overlap score with the reference instances (as measured by ROUGE-L or BLEURT) is statistically significantly better with the completions from guided instruction compared to a "general instruction" that does not include the dataset and partition name. The second idea marks a dataset partition as contaminated if a classifier based on GPT-4 with few-shot in-context learning prompt marks multiple generated completions as exact/near-exact matches of the corresponding reference instances. Our best method achieves an accuracy between 92% and 100% in detecting if an LLM is contaminated with seven datasets, containing train and test/validation partitions, when contrasted with manual evaluation by human experts. Further, our findings indicate that GPT-4 is contaminated with AG News, WNLI, and XSum datasets.
Forward citations
Cited by 18 Pith papers
-
Fully Open Meditron: An Auditable Pipeline for Clinical LLMs
Presents the first fully open pipeline for clinical LLMs by unifying eight public QA datasets with three clinician-vetted synthetic extensions and applying it to five base models to achieve benchmark gains while maint...
-
WARP: Weight-Space Analysis for Recovering Training Data Portfolios
WARP recovers training domain mixtures from fine-tuned model weights using weight-space interpolation via model merging to generate pseudo-checkpoints and geometric features mapped to proportions.
-
AlgoBench: Benchmarking Algorithmic Adaptation in Code Generation
AlgoBench creates traceable variants of competitive programming problems via constraint shifts that invalidate original algorithms, paired with complexity metrics that reveal LLMs often produce functionally correct bu...
-
Fully Open Meditron: An Auditable Pipeline for Clinical LLMs
Presents the first fully open pipeline for clinical LLMs that unifies eight public QA datasets with clinician-vetted synthetic data from guidelines and vignettes, achieving improved performance on medical benchmarks w...
-
Detecting Pretraining Data from Large Language Models
Min-K% Prob detects pretraining data in LLMs by flagging outlier low-probability words in text, achieving 7.4% better performance than prior methods on the new WIKIMIA benchmark.
-
Representational Depth of Evaluation Awareness Shifts With Scale in Open-Weight Language Models
In Qwen 2.5 and Gemma 2 families, the layer where evaluation awareness is most linearly recoverable shifts from late layers in small models to early layers in large models.
-
The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection
Systematic tests of LLM contamination detectors across 27 models show frequent failures from distribution shift and scale, concluding statistical methods cannot replace transparent data provenance.
-
The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection
Empirical evaluation across 25 LLMs shows contamination detection methods achieve correct outcomes in only 201 of 335 cases, exposing failure modes from distribution shift and benchmark scale.
-
Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents
Memory-equipped LLM agents exhibit increasing safety violation rates as memory accumulates across independent tasks, termed temporal memory contamination, detected via a new trigger-probe protocol.
-
iCoRe: An Iterative Correlation-Aware Retriever for Bug Reproduction Test Generation
iCoRe improves Fail-to-Pass rates to 42.0% and 52.8% on two bug reproduction benchmarks by using correlation-aware iterative retrieval instead of standard semantic or BM25 methods.
-
iCoRe: An Iterative Correlation-Aware Retriever for Bug Reproduction Test Generation
iCoRe improves bug reproduction test generation by combining differentiated code/test retrieval, function-call-structure similarity, and iterative generation-to-retrieval feedback, achieving state-of-the-art results o...
-
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
-
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
Self-RAG trains LLMs to adaptively retrieve passages on demand and self-critique using reflection tokens, outperforming ChatGPT and retrieval-augmented Llama2 on QA, reasoning, and fact verification.
-
The Illusion of Reasoning: Exposing Evasive Data Contamination in LLMs via Zero-CoT Truncation
ZCP detects direct and evasive data contamination in LLMs by truncating CoT reasoning and contrasting zero-CoT accuracy on original versus perturbed isomorphic datasets, plus a Contamination Confidence metric.
-
Interactive Evaluation Requires a Design Science
Interactive evaluation of AI must be reframed as a distinct paradigm that maps interaction trajectories to judgments on process, recoverability, coordination, robustness, and system performance, supported by a two-axi...
-
Revisiting the Past: Data Unlearning with Model State History
MSA performs data unlearning in LLMs by arithmetic operations on prior model checkpoints to remove targeted datapoint influence, with experiments showing competitive or better results than existing unlearning methods.
-
CodePori: Large-Scale System for Autonomous Software Development Using Multi-Agent Technology
CodePori is a multi-agent LLM system for code generation whose participant evaluation identifies practical challenges like memory limits and hallucinations missed by binary benchmarks.
-
The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Survey
A survey of emerging AI agent architectures that organizes single and multi-agent designs around reasoning, planning, tool use, communication, and reflection phases.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.