mlsanityData to 13 Oct 2026

Right Knowledge, Wrong Answer: Characterizing Parametric Temporal Conflict in Open-Weight Language Models

arXiv 2606.20959 ↗First posted 18 Jun 2026cs.LG cs.CLcited by 2

Not on the map. Why?

Abstract (arXiv)

Language models may encode both outdated facts and their newer replacements. We introduce Parametric Temporal Conflict (PTC), where the newer fact is present and recoverable, but the default forward pass prefers the outdated one. We release a deterministically verified benchmark of 8,746 Wikidata position-holder transitions and evaluate four open-weight language models across three families. A date-prefix prompt recovers the newer fact in 61-81% of PTC cases. Activation patching flips predictions in 72-85% of cases and localizes the preference to model-specific upper-layer regions. Residual-stream steering outperforms norm-matched random directions, indicating direction-specific representations. These results show that PTC reflects a localized representational preference rather than missing knowledge. Recovery is measured on oracle-identified conflicts because automatic detection remains unreliable. We release the benchmark, code, and statistics.

Key facts. Cited by 2 papers in the map (data through 13 Oct 2026). Builds on 29 of its 36 references that are in the map. First posted on arXiv 18 Jun 2026.

Builds on

29 of 36 references.

References · 29

  1. 2022Locating and Editing Factual Associations in GPTcited by 2,650
  2. 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,323
  3. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,280
  4. 2022Mass-Editing Memory in a Transformercited by 990
  5. 2023FreshLLMs: Refreshing Large Language Models with Search Engine Augmentationcited by 312
  6. 2021Time-Aware Language Models as Temporal Knowledge Basescited by 291
  7. 2024Knowledge Conflicts for LLMs: A Surveycited by 277
  8. 2022RealTime QA: What's the Answer Right Now?cited by 242
  9. 2023Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Modelscited by 240
  10. 2021A Dataset for Answering Time-Sensitive Questionscited by 152
  11. 2021SituatedQA: Incorporating Extra-Linguistic Contexts into QAcited by 145
  12. 2024Test of Time: A Benchmark for Evaluating LLMs on Temporal Reasoningcited by 84
17 more
  1. 2024DYNAMICQA: Tracing Internal Knowledge Conflicts in Language Modelscited by 34
  2. 2023Time is Encoded in the Weights of Finetuned Language Modelscited by 31
  3. 2025Taming Knowledge Conflicts in Language Modelscited by 28
  4. 2023Towards Robust Temporal Reasoning of Large Language Models via a Multi-Hop QA Dataset and Pseudo-Instruction Tuningcited by 18
  5. 2023Carpe Diem: On the Evaluation of World Knowledge in Lifelong Language Modelscited by 17
  6. 2024UnSeenTimeQA: Time-Sensitive Question-Answering Beyond LLMs' Memorizationcited by 17
  7. 2023Question Answering as Programming for Solving Time-Sensitive Questionscited by 14
  8. 2024Time Awareness in Large Language Models: Benchmarking Fact Recall Across Timecited by 12
  9. 2026Cornerstones or Stumbling Blocks? Deciphering the Rock Tokens in On-Policy Distillationcited by 11
  10. 2024MuLan: A Study of Fact Mutability in Language Modelscited by 10
  11. 2025Model Whisper: Steering Vectors Unlock Large Language Models' Potential in Test-timecited by 7
  12. 2026Omni-Modal Dissonance Benchmark: Systematically Breaking Modality Consensus to Probe Robustness and Calibrated Abstentioncited by 6
  13. 2026TRIAGE: Evaluating Prospective Metacognitive Control in LLMs under Resource Constraintscited by 6
  14. 2026Where Knowledge Collides: A Mechanistic Study of Intra-Memory Knowledge Conflict in Language Modelscited by 5
  15. 2025Temporal Alignment of Time Sensitive Facts with Activation Engineeringcited by 4
  16. 2026UAT-LITE: Inference-Time Uncertainty-Aware Attention for Pretrained Transformerscited by 4
  17. 2026Temporal Fact Conflicts in LLMs: Reproducibility Insights from Unifying DYNAMICQA and MULANcited by 2

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Reachability Is Not Realization: Tracing the Sources of LLM Benchmark Gainscited by 1
  2. 2026New Evidence, Same Choice: Testing Physical Experiment Selection in Vision Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts
  2. 2026Large Language Models in Resolving Contextual Knowledge Conflicts
  3. 2025Conflict-Aware Fusion: Mitigating Logic Inertia in Large Language Models via Structured Cognitive Priors
  4. 2026Seeing through the Conflict: Transparent Knowledge Conflict Handling in Retrieval-Augmented Generation
  5. 2026MemToC: Benchmarking Memory-Tool Conflict Resolution in Large Language Models
  6. 2026Simulated Adoption: Decoupling Magnitude and Direction in LLM In-Context Conflict Resolution
  7. 2025FaithfulRAG: Fact-Level Conflict Modeling for Context-Faithful Retrieval-Augmented Generation
  8. 2026STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?