ARXIV:2604.19884 · LLM OPTIMIZATION · SUBMITTED 23 APR · 05:10 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedVerifiedPaperPack: citation fields availablePartialProof: unverified proof status

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

Chenxi Zhou · Pengfei Cao · Jiang Li · Bohan Yu · Jinyu Ye · Jun Zhao · +1 at arXiv

Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

Ship in 2-4 weeks›Score3.0Evidence unverified

Opportunity summary

Pain Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

Evidence 0 refs | 3 sources | 50% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency. While 4-bit quantization is widely regarded as an optimal trade-off, reducing the precision to 2-bit usually triggers a catastrophic ``performance…

METHOD

Full abstract

Post-Training Quantization (PTQ) is critical for the efficient deployment of Large Language Models (LLMs). While 4-bit quantization is widely regarded as an optimal trade-off, reducing the precision to 2-bit usually triggers a catastrophic ``performance cliff.'' It remains unclear whether the underlying mechanisms differ fundamentally. Consequently, we conduct a systematic mechanistic analysis, revealing two qualitatively distinct failure modes: Signal Degradation, where the computational patterns remain intact but information precision is impaired by cumulative error; and Computation Collapse, where key components fail to function, preventing correct information processing and destroying the signal in the early layers. Guided by this diagnosis, we conduct mechanism-aware interventions, demonstrating that targeted, training-free repair can mitigate Signal Degradation, but remains ineffective for Computation Collapse. Our findings provide a systematic diagnostic framework for PTQ failures and suggest that addressing Computation Collapse requires structural reconstruction rather than mere compensation.

RESULT

ScienceToStartup currently rates this 3.0/10 on the public viability pass. Our findings provide a systematic diagnostic framework for PTQ failures and suggest that addressing Computation Collapse requires structural reconstruction rather than mere compensation. Code…

WHY NOW

LLM Optimization moved forward this cycle; last verified April 2026. Public score 3.0/10. Production flags indicate code availability.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score3.0

PainUncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

Evidence0 refs | 3 sources | 50% coverage

Blockerno shell-level blocker reported

Analysis summary

Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

VerifiedSource: PDF linkedVerifiedPaperPack: citation fields availablePartialProof: unverified proof status

Competitive landscape

Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

Segment

LLM Optimization

Adoption evidence

No public code link in the paper record yet

Commercial read

3.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

{ "contract_version": "paper-r2", "paper_id": "b724b39b-5831-43de-8394-2509c8c2b648", "arxiv_id": "2604.19884", "canonical_route": "/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "active_tab": "synced from current hash by the drawer client", "selected_artifact": "from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "endpoints": { "paper_pack": "/api/v1/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization/paper-pack", "build_passport": "/api/v1/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization/build-passport", "mcp_resource": "sciencetostartup://surfaces/paper-workspace" } }

{ "surface": "paper", "mode": "paper", "query": "From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization", "normalized_query": "2604.19884", "route": "/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "paper_ref": "from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "topic_slug": null, "benchmark_ref": null, "dataset_ref": null }

{ "@context": "https://schema.org", "@graph": [ { "@type": "WebPage", "@id": "https://sciencetostartup.com/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization#webpage", "url": "https://sciencetostartup.com/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "name": "From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization", "description": "Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.", "isPartOf": { "@id": "https://sciencetostartup.com/#website" } }, { "@type": "ScholarlyArticle", "@id": "https://sciencetostartup.com/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization#scholarlyArticle", "headline": "From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization", "description": "Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.", "url": "https://sciencetostartup.com/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization", "sameAs": "https://arxiv.org/abs/2604.19884", "identifier": { "@type": "PropertyValue", "propertyID": "arXiv", "value": "2604.19884" }, "isAccessibleForFree": true, "isPartOf": { "@id": "https://sciencetostartup.com/#website" }, "datePublished": "2026-04-21T18:03:38.000Z", "author": [ { "@type": "Person", "name": "Chenxi Zhou" }, { "@type": "Person", "name": "Pengfei Cao" }, { "@type": "Person", "name": "Jiang Li" }, { "@type": "Person", "name": "Bohan Yu" }, { "@type": "Person", "name": "Jinyu Ye" }, { "@type": "Person", "name": "Jun Zhao" }, { "@type": "Person", "name": "Kang Liu" } ], "additionalProperty": [ { "@type": "PropertyValue", "propertyID": "viabilityScore", "value": 3 }, { "@type": "PropertyValue", "propertyID": "researchDomain", "value": "LLM Optimization" }, { "@type": "PropertyValue", "propertyID": "commercialReadiness", "value": "code" } ] }, { "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://sciencetostartup.com" }, { "@type": "ListItem", "position": 2, "name": "LLM Optimization", "item": "https://sciencetostartup.com/topics" }, { "@type": "ListItem", "position": 3, "name": "From Signal Degradation to Computation Collapse: Uncovering ", "item": "https://sciencetostartup.com/paper/from-signal-degradation-to-computation-collapse-uncovering-the-two-failure-modes-of-llm-quantization" } ] } ] }

Competitive landscape

Uncovering and diagnosing two distinct failure modes in LLM quantization to improve model efficiency.

Segment

LLM Optimization

Adoption evidence

No public code link in the paper record yet

Commercial read

3.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline