Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign inputs. We propose \\textbf{C}ontinuous \\textbf{L}at\\textbf{E}nt \\textbf{A}dapter \\textbf{R}outing (CLEAR), a conditional safety adaptation framework that uses a lightweight hidden-state gate to continuously control the activation strength of a safety low-rank adapter. CLEAR aims to reduce harmful completions while avoiding unnecessary changes to the frozen backbone that could degrade performance on benign prompts. Experiments on widely used safety and utility benchmarks show that CLEAR improves robustness on HarmBench while reducing the utility degradation observed with globally applied safety tuning such as SFT or standard low-rank adaptation (LoRA). On Llama-3-8B-Instruct, CLEAR reduces HarmBench ASR from 32.3% to 0.5%, while retaining most of the base model's utility and achieving up to 7.1 percentage points higher GSM8K accuracy than globally applied SFT or LoRA. These results suggest that CLEAR is a promising mechanism for improving the safety--utility trade-off in LLM alignment.</p>\n","updatedAt":"2026-08-24T02:02:00.390Z","author":{"_id":"6058351db2c84d5386b3afe5","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6058351db2c84d5386b3afe5/mS1Nu9i5VYgAnHSKXz-b4.jpeg","fullname":"Enyi (Olivia) Jiang","name":"EnyiJiang","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8972113728523254},"editors":["EnyiJiang"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/6058351db2c84d5386b3afe5/mS1Nu9i5VYgAnHSKXz-b4.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.21278","authors":[{"_id":"6a8ba5af3d26296ea30918f5","name":"Chengxiao Wang","hidden":false},{"_id":"6a8ba5af3d26296ea30918f6","name":"Enyi Jiang","hidden":false},{"_id":"6a8ba5af3d26296ea30918f7","name":"Xiaojing Liao","hidden":false},{"_id":"6a8ba5af3d26296ea30918f8","name":"Sanmi Koyejo","hidden":false}],"publishedAt":"2026-08-21T00:00:00.000Z","submittedOnDailyAt":"2026-08-24T00:00:00.000Z","title":"CLEAR: Continuous Latent Adapter Routing for Utility-Preserving LLM Safety Alignment","submittedOnDailyBy":{"_id":"6058351db2c84d5386b3afe5","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6058351db2c84d5386b3afe5/mS1Nu9i5VYgAnHSKXz-b4.jpeg","isPro":true,"fullname":"Enyi (Olivia) Jiang","user":"EnyiJiang","type":"user","name":"EnyiJiang"},"summary":"Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign inputs. We propose Continuous LatEnt Adapter Routing (CLEAR), a conditional safety adaptation framework that uses a lightweight hidden-state gate to continuously control the activation strength of a safety low-rank adapter. CLEAR aims to reduce harmful completions while avoiding unnecessary changes to the frozen backbone that could degrade performance on benign prompts. Experiments on widely used safety and utility benchmarks show that CLEAR improves robustness on HarmBench while reducing the utility degradation observed with globally applied safety tuning such as SFT or standard low-rank adaptation (LoRA). On Llama-3-8B-Instruct, CLEAR reduces HarmBench ASR from 32.3\\% to 0.5\\%, while retaining most of the base model's utility and achieving up to 7.1 percentage points higher GSM8K accuracy than globally applied SFT or LoRA. These results suggest that CLEAR is a promising mechanism for improving the safety--utility trade-off in LLM alignment.","upvotes":1,"discussionId":"6a8ba5af3d26296ea30918f9","ai_summary":"CLEAR uses a hidden-state gate to continuously modulate a safety low-rank adapter, improving LLM safety while preserving utility on benign inputs.","ai_keywords":["low-rank adapter","hidden-state gate","conditional safety adaptation","CLEAR","LoRA","HarmBench","LLM alignment"],"ai_summary_model":"thinkingmachines/Inkling-Small","organization":{"_id":"65448bef5b5d9185ba3202b9","name":"UIUC-CS","fullname":"University of Illinois at Urbana-Champaign","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/65448b21fcb96b8b48733729/ycqcXFayMTTD_KpE37067.jpeg"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"63ac5701c21e60a3e9b58aa7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ac5701c21e60a3e9b58aa7/g6EX7diOpuA94R2ab-rZC.png","isPro":true,"fullname":"Dipankar Sarkar","user":"dipankarsarkar","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"65448bef5b5d9185ba3202b9","name":"UIUC-CS","fullname":"University of Illinois at Urbana-Champaign","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/65448b21fcb96b8b48733729/ycqcXFayMTTD_KpE37067.jpeg"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.21278.md","query":{}}">
CLEAR: Continuous Latent Adapter Routing for Utility-Preserving LLM Safety Alignment
Abstract
CLEAR uses a hidden-state gate to continuously modulate a safety low-rank adapter, improving LLM safety while preserving utility on benign inputs.
Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign inputs. We propose Continuous LatEnt Adapter Routing (CLEAR), a conditional safety adaptation framework that uses a lightweight hidden-state gate to continuously control the activation strength of a safety low-rank adapter. CLEAR aims to reduce harmful completions while avoiding unnecessary changes to the frozen backbone that could degrade performance on benign prompts. Experiments on widely used safety and utility benchmarks show that CLEAR improves robustness on HarmBench while reducing the utility degradation observed with globally applied safety tuning such as SFT or standard low-rank adaptation (LoRA). On Llama-3-8B-Instruct, CLEAR reduces HarmBench ASR from 32.3\% to 0.5\%, while retaining most of the base model's utility and achieving up to 7.1 percentage points higher GSM8K accuracy than globally applied SFT or LoRA. These results suggest that CLEAR is a promising mechanism for improving the safety--utility trade-off in LLM alignment.
Community
Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign inputs. We propose \textbf{C}ontinuous \textbf{L}at\textbf{E}nt \textbf{A}dapter \textbf{R}outing (CLEAR), a conditional safety adaptation framework that uses a lightweight hidden-state gate to continuously control the activation strength of a safety low-rank adapter. CLEAR aims to reduce harmful completions while avoiding unnecessary changes to the frozen backbone that could degrade performance on benign prompts. Experiments on widely used safety and utility benchmarks show that CLEAR improves robustness on HarmBench while reducing the utility degradation observed with globally applied safety tuning such as SFT or standard low-rank adaptation (LoRA). On Llama-3-8B-Instruct, CLEAR reduces HarmBench ASR from 32.3% to 0.5%, while retaining most of the base model's utility and achieving up to 7.1 percentage points higher GSM8K accuracy than globally applied SFT or LoRA. These results suggest that CLEAR is a promising mechanism for improving the safety--utility trade-off in LLM alignment.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.21278 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.21278 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.21278 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.