Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

CeRA: Breaking the Linear Ceiling of Low-Rank Adaptation with Non-linearity Retained at Inference

Hung-Hsuan Chen

Published

Feb 26, 2026

Citations

0

Trust level

Low

Usefulness score

5/100 (Low)

Extraction confidence

45% (Low)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

Jun 25, 2026

Open arXiv Find implementation Shortlist (0) RSS feed

Tags

Automatic Metrics

Should you rely on this paper?

This paper is adjacent to HFEPX scope and is best used for background context, not as a primary protocol reference.

Use this as background context only. Do not make protocol decisions from this page alone.

Best use

Background context only

Use if you need

A benchmark-and-metrics comparison anchor.

What to verify

Validate the evaluation procedure and quality controls in the full paper before operational use.

Main weakness

This paper looks adjacent to evaluation work, but not like a strong protocol reference.

Human feedback signal
Not explicit
Not explicit in abstract metadata
Evaluation signal
Detected
Eval setup described
Usefulness for eval research
5/100
Adjacent candidate

Treat as adjacent context, not a core eval-method reference.

If you are doing eval pipeline work, start here

Human Eval Papers LLM-as-Judge Papers Pairwise Preference Papers Tool-Use Evaluation Search

Abstract

Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT). However, it faces a ``linear ceiling'': increasing the rank yields diminishing returns in expressive capacity due to linear constraints. We introduce CeRA (Capacity-enhanced Rank Adaptation), a weight-level parallel adapter that injects SiLU gating and dropout to induce non-linearity during inference, thereby placing it in a different function class from adapters whose non-linearity exists during training and collapses to an affine map at inference time. On both the basic arithmetic (GSM8K) and the complex MATH benchmark, CeRA is markedly more parameter-efficient. Across a full rank $\times$ learning rate sweep, CeRA at rank 64 achieves the highest MATH pass@1 of any configuration in the grid (23.6\%), matching or exceeding both a rank-512 LoRA (22.4\%) and DoRA (19.8\%) while using only 1/8 of the parameter budget. With the rank and learning rate fixed, CeRA equals or outperforms LoRA in 10 of 12 matched settings. Spectral analysis attributes the gain, at least in part, to smooth (SiLU) gating, which broadens the utilization of the singular-value spectrum and mitigates the rank collapse that linear adapters exhibit at high rank. Additionally, dropout appears to contribute to regularization rather than rank expansion.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

missing

None explicit

No explicit feedback protocol extracted.

"Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT)."

Evaluation Modes

partial

Automatic Metrics

Includes extracted eval setup.

"Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT)."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT)."

Benchmarks / Datasets

partial

GSM8K

Useful for quick benchmark comparison.

"On both the basic arithmetic (GSM8K) and the complex MATH benchmark, CeRA is markedly more parameter-efficient."

Reported Metrics

partial

Accuracy, Perplexity, Pass@1

Useful for evaluation criteria comparison.

"Across a full rank $\times$ learning rate sweep, CeRA at rank 64 achieves the highest MATH pass@1 of any configuration in the grid (23.6\%), matching or exceeding both a rank-512 LoRA (22.4\%) and DoRA (19.8\%) while using only 1/8 of the parameter budget."

Benchmarks and datasets

GSM8K

Reported metrics

accuracyperplexitypass@1
Human feedback details
Uses human feedback
No
Feedback types
None
Rater population
Not reported
Expertise required
Math
Evaluation details
Evaluation modes
Automatic Metrics
Agentic eval
None
Quality controls
Not reported
Evidence quality
Low
Use this page as
Background context only

Research brief

Metadata summary

Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT).

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT).
  • However, it faces a ``linear ceiling'': increasing the rank yields diminishing returns in expressive capacity due to linear constraints.
  • We introduce CeRA (Capacity-enhanced Rank Adaptation), a weight-level parallel adapter that injects SiLU gating and dropout to induce non-linearity during inference, thereby placing it in a different function class from adapters whose non-linearity exists during training and collapses to an affine map at inference time.

Researcher actions

  • Compare this paper against others mentioning GSM8K.
  • Check the full text for explicit evaluation design choices (raters, protocol, and metrics).
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Recommended queries

GSM8K

Contribution summary

  • We introduce CeRA (Capacity-enhanced Rank Adaptation), a weight-level parallel adapter that injects SiLU gating and structural dropout to induce manifold expansion.
  • On the SlimOrca benchmark, CeRA breaks this linear barrier: at rank 64 (PPL 3.89), it outperforms LoRA at rank 512 (PPL 3.90), demonstrating superior spectral efficiency.
  • Remarkably, CeRA at rank 64 (Pass@1 16.36\%) outperforms LoRA at rank 512 (Pass@1 15.72\%), achieving superior reasoning accuracy with only 1/8 of the parameter budget.

Why it matters for eval

  • On the SlimOrca benchmark, CeRA breaks this linear barrier: at rank 64 (PPL 3.89), it outperforms LoRA at rank 512 (PPL 3.90), demonstrating superior spectral efficiency.

Researcher checklist

  • Human feedback protocol is explicit

    No explicit human feedback protocol detected.

  • Evaluation mode is explicit

    Detected: Automatic Metrics

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    Detected: GSM8K

  • Metric reporting is present

    Detected: accuracy, perplexity, pass@1