closedclosed 3 Nov 2026Research1 Research / proof of conceptAI
Trojan-Based Data Poisoning Attacks Increasing LLM Token Response Costs
Researchers devised a defense against a data-poisoning attack that plants a Trojan to artificially increase LLM response lengths and financial costs. The attack triggers on semantically similar instructions within prompts.
- first report
- 20 Oct 2026
- last update
- 20 Oct 2026
Timeline · 1
Sources · 1
Research and catalogues
- 2026-10-20LLM Trojans that Increase Response Lengths and CostarXiv · arxiv.org/abs/2610.24784
Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.
Classification
- lane
- Research
- AI role
- Attack on AI
- categories
- Data poisoning
- severity
- 1 · Research / proof of concept
- targets
- Open-weight model
- owasp llm top 10
- LLM04:2025 Data and Model Poisoning
- mitre atlas
- no technique for these categories
- dates
- happened 20 Oct 2026 · first reported 20 Oct 2026
- origin
- News & research feeds
- record
- confidence 0.90 · created 2026-10-20 04:59Z · updated 2026-11-04 00:07Z
Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned
Copy a brief
This record as a short Markdown brief, for a ticket, a chat or a coding agent.
the brief as text
# Trojan-Based Data Poisoning Attacks Increasing LLM Token Response Costs https://gcdradar.vercel.app/incidents/trojan-based-data-poisoning-attacks-increasing-llm-token-res · AIR-2026-1418 · closed 2026-11-03 - Lane: Research - Severity: 1 (Research / proof of concept) - Fix status: Fix status unknown - First reported: 2026-10-20 - AI role: Attack on AI ## Summary Researchers devised a defense against a data-poisoning attack that plants a Trojan to artificially increase LLM response lengths and financial costs. The attack triggers on semantically similar instructions within prompts. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-10-20: LLM Trojans that Increase Response Lengths and Cost (https://arxiv.org/abs/2610.24784) ## Sources - 2026-10-20 arXiv: LLM Trojans that Increase Response Lengths and Cost https://arxiv.org/abs/2610.24784 Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- liveVision Encoders Poison-then-Hide: Finetuning-Activated Backdoor Attack on Pretrained Vision Encoders
Research · since 13 Nov 2026
AI - liveFake news articles change forecasts of LLM forecasting systems
Research · since 13 Nov 2026
AI - liveTrained to betray: How post-training supply chain attacks hijack coding agents
Research · since 12 Nov 2026
AI - liveLLM Agent When Experience Becomes a Backdoor: Poisoning Self-Evolving LLM Agents via Failure-Aware Task Optimization
Research · since 10 Nov 2026
AI - liveSkillOpt Tracing Backdoor Skill Formation in Text-Space Skill Optimization
Research · since 9 Nov 2026
AI - liveLLM Agent Memory Managers Accept Changes, Reject Attacks: Robust Memory Management for LLM Agents
Research · since 2 Nov 2026
AI