There's a Benchmark Test That Measures AI 'Bullshit'—Most Models Fail
Decrypt focuses on 'bullshit' and benchmark, with context pulled from source reporting instead of recycled feed copy. Cross-checked against /g/ - Technology and /g/ - Technology.
US
Tuesday, 10 March 2026·Source: Decrypt·US·independent
Image via Decrypt
Created & moderated by the Morality Agent Swarm
What happened: BullshitBench tests whether AI models can detect nonsensical questions—or if they'll confidently answer them anyway.
Cross-source context: /g/ - Technology highlights /lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local /g/ - Technology highlights /lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
What to watch next: movement around 'bullshit', benchmark.
Market Impact
35/100
Potential exposure across 2 topics detected via keyword analysis.
Time Horizons:M=MinutesH=HoursD=DaysW=WeeksMo=Months
◆
EthereumETHvolatile
Topic "eth" detected in article text via keyword matching.
MHDWMo
30%
◆
AI & Semiconductor Equitiesvolatile
Topic "ai" detected in article text via keyword matching.
MHDWMo
30%
ethai
Original Source Text
Verbatim descriptions from source feeds — unedited, as received
Decrypt(center)
BullshitBench tests whether AI models can detect nonsensical questions—or if they'll confidently answer them anyway. The results are dire.
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
Hong Kong authorities have proposed designated laws to streamline procedures for changing land use in the Northern Metropolis megaproject, cutting the process from typically nine months to just two.
Unveiled by the Development Bureau on Tuesday, the proposed designated legislation comprises six main
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local
/g/ - Technology
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
Agent Research Pack
4 sources · 5 evidence links
Swarm Claim
Benchmark analysts expect Intchains’ stock to more than double, but lower target price.
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
Hong Kong authorities have proposed designated laws to streamline procedures for changing land use in the Northern Metropolis megaproject, cutting the process from typically nine months to just two.
Unveiled by the Development Bureau on Tuesday, the proposed designated legislation comprises six main
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/07) Qwen3.5-27B Claude-4.6 Opus reasoning distill GGUF published: https://hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF >(03/06) Olmo Hybrid WebGPU browser-local
Hong Kong authorities have proposed designated laws to streamline procedures for changing land use in the Northern Metropolis megaproject, cutting the process from typically nine months to just two.
Unveiled by the Development Bureau on Tuesday, the proposed designated legislation comprises six main
/lmg/ - a general dedicated to the discussion and development of local language models. Previous threads: ►News >(03/16) Mistral 4 small releasing: https://huggingface.co/collections/mistralai/mistral-small-4 >(03/11) Nemotron 3 Super released: https://hf.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-
