Quantization Amplifies Determinism, Not Bias: Scale-Dependent Behavioral Effects of Serving-Time Weight Compression

• URL: http://arxiv.org/abs/2609.07901v1
• Date: Mon, 07 Sep 2026 19:06:31 GMT
• Title: Quantization Amplifies Determinism, Not Bias: Scale-Dependent Behavioral Effects of Serving-Time Weight Compression
• Abstract summary: quantization of mid-sized models is often considered "nearly free"<n>We examine a different question: when several answers are valid, does quantization change what a model chooses to say?<n>We serve three checkpoints (Qwen3-8B/14B/32B) at three weight precisions (W4A16 AWQ, W8A16 FP8-Marlin, and bf16)<n>We collect approximately 71,000 completions paired by prompt and seed across two custom, leak-checked prompt batteries.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Weight quantization largely determines the economics of serving open-weight LLMs. Its costs are usually assessed with capability benchmarks, on which 4-bit quantization of mid-sized models is often considered "nearly free." We examine a different question: when several answers are valid, does quantization change what a model chooses to say? We serve three checkpoints (Qwen3-8B/14B/32B) at three weight precisions (W4A16 AWQ, W8A16 FP8-Marlin, and bf16), holding the hardware, software, and sampling configuration constant, and collect approximately 71,000 completions paired by prompt and seed across two custom, leak-checked prompt batteries. We pre-specified the analyses in three waves in version control. At 8B, int4 reduces output diversity: the probability that two samples for the same scenario recommend the same brand increases by 5.1 percentage points (prompt-paired sign-flip test, Holm p = .023; reproduced at +4.4pp on a full regeneration of the arm), and lexical diversity falls substantially (TTR -0.011, standardized effect -0.51; robust to a length-controlled measure). At 14B and 32B, no content-concentration measure reaches significance; instead, stylistic drift emerges (em-dash rate +0.46/1k words at 14B and +0.61/1k at 32B, both Holm p <= .0024). Pre-specified tests of stereotype direction are null at every scale: outputs concentrate on the modal answer for each prompt rather than on stereotypical answers. Mechanistically, the token-level distribution becomes flatter (decision-token entropy +0.091 bits, p = .015) while the semantic distribution, measured directly from first-token log probabilities, becomes more concentrated (collision +2.6pp, p = .023): individual tokens become less predictable even as meanings become more repetitive. At 8B, the smallest size tested, AWQ-int4 serving measurably narrows the range of suggestions; audits should assess concentration as well as bias.

Related papers

• The Dice Roll Method: A Standardized Protocol for Repeated-Query Auditing of Large Language Model Brand Recommendations [0.0]
  We formalize the Dice Roll Method as a reusable protocol for repeated-query auditing of brand recommendations.<n>Three tiers of iteration guidance emerge from the D-study: exploratory, confirmatory, and rigorous.<n>A pre-registered external validation on three independent corpora reproduces the D-study reliability prediction in 37 of 39 cells with no failures.
  arXiv  Detail & Related papers  (2026-09-03T16:22:06Z)
• Deterministic LLM Inference Across GPU Kernels: Power-of-Two INT8 Quantization Scales and the Limits of Tolerance-Based Conformance [0.10152838128195468]
  Conformance suites for quantized GEMM kernels ask whether two implementations agree within a tolerance.<n>We find that every one of five epilogue faults -- scale precision, double rounding, multiplication order, output truncation, fused ordering -- moves the output by at most a single bfloat16 spacing.<n>Four of the five faults are detected by no check in the suite, and the fifth only under power-of-two scales.
  arXiv  Detail & Related papers  (2026-08-25T03:05:54Z)
• RDQ: Residual Distribution Quantization for Large Language Models [3.7323766833175718]
  Post-training quantization of large language models degrades sharply below 4-bit precision.<n>We propose RDQ (Residual Distribution Quantization), a PTQ framework whose central contribution is Cascaded Error Compensation.<n>All output is standard group-128 asymmetric quantization, deployable on Qualcomm AIMET, GGUF, and any standard inference stack at zero runtime overhead.
  arXiv  Detail & Related papers  (2026-07-11T05:54:00Z)
• Silent Failures in Quantized LLM Reasoning: A Taxonomy-Based Analysis of Hollow Convergence and Failure Mode Shifts [0.0]
  We show that post-training quantization can silently alter how large language models reason even when task accuracy is preserved.<n>We classify 30,000 chain-of-thought outputs from five instruction-tuned LLMs across three quantization precisions.<n>We find that while accuracy is robust across precisions, Hollow Convergence shows a significant size-dependent shift under NF4.
  arXiv  Detail & Related papers  (2026-07-10T21:55:05Z)
• Automated Proving of Shannon-Type Entropy Inequalities via Fine-Tuned Language Models and Guided Tree Search [50.16356451328644]
  Proving Shannon-type entropy inequalities is a fundamental task in information theory.<n>We investigate whether small-scale large language models, fine-tuned on atomic proof steps, can automate this process.<n>We show that GPT-5.5 solves 1.7% samples under zero-shot prompting while Psitip solves 33.3% samples.
  arXiv  Detail & Related papers  (2026-06-04T05:43:12Z)
• Right Makes Might: Aligning Verified Hidden States Empowers RL Reasoning [55.264863369127774]
  Current methods reduce each correct rollout to a single reward bit, ignoring the geometric structure shared among their hidden states.<n>We propose Hidden-Align, an auxiliary loss function that aligns the last-layer hidden states of correct rollouts at the anchor token during RL training, with zero overhead in both training and inference.<n>On eight mathematical reasoning benchmarks, Hidden-Align improves average pass@1 over the DAPO baseline by 3.8, 6.2, and 5.4 percentage points on Qwen3-1.7B, 4B, and 14B respectively, with consistent pass@k gains across all three scales.
  arXiv  Detail & Related papers  (2026-06-02T06:51:15Z)
• The Last Word Often Wins: A Format Confound in Chain-of-Thought Corruption Studies [0.0]
  Corruption studies infer which steps are computationally important'' from accuracy loss when steps are corrupted.<n>We show that when benchmark chains end with an explicit terminal answer line, these tests largely measure emphanswer placement rather than where intermediate computation is carried out.
  arXiv  Detail & Related papers  (2026-05-11T16:26:50Z)
• Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels [0.0]
  Large Language Models are routinely compressed via post-training quantization to reduce inference costs and memory footprint for cloud and edge deployment.<n>Existing studies typically compare only two conditions, rely on aggregate bias metrics, and evaluate a single model family.<n>We conduct a controlled empirical study of three instruction-tuned models at five precision levels (BF16 through 3-bit) on 12,148 BBQ bias benchmark items.
  arXiv  Detail & Related papers  (2026-05-02T05:41:47Z)
• ParetoQ: Improving Scaling Laws in Extremely Low-bit LLM Quantization [73.60493264901359]
  We present a unified framework for rigorous comparisons across 1-bit, 1.58-bit, 2-bit, 3-bit, and 4-bit quantization settings.<n>We show that ternary, 2-bit, and 3-bit quantization maintains comparable performance in the size-accuracy trade-off.<n>Considering hardware constraints, 2-bit quantization offers promising potential for memory reduction and speedup.
  arXiv  Detail & Related papers  (2025-02-04T18:59:26Z)
• "Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization [67.3213104337679]
  Quantization is a powerful tool for accelerating large language model (LLM) inference, but the accuracy-performance trade-offs across different formats remain unclear.<n>We conduct the most comprehensive empirical study to date, evaluating FP8, INT8, and INT4 quantization across academic benchmarks and real-world tasks.
  arXiv  Detail & Related papers  (2024-11-04T18:21:59Z)
• Large Language Monkeys: Scaling Inference Compute with Repeated Sampling [81.34900892130929]
  We explore inference compute as another axis for scaling, using the simple technique of repeatedly sampling candidate solutions from a model.<n>Across multiple tasks and models, we observe that coverage scales with the number of samples over four orders of magnitude.<n>In domains like coding and formal proofs, where answers can be automatically verified, these increases in coverage directly translate into improved performance.
  arXiv  Detail & Related papers  (2024-07-31T17:57:25Z)
• The case for 4-bit precision: k-bit Inference Scaling Laws [75.4335600212427]
  Quantization methods reduce the number of bits required to represent each parameter in a model. The final model size depends on both the number of parameters of the original model and the rate of compression. We run more than 35,000 zero-shot experiments with 16-bit inputs and k-bit parameters to examine which quantization methods improve scaling for 3 to 8-bit precision.
  arXiv  Detail & Related papers  (2022-12-19T18:48:33Z)
• AvgOut: A Simple Output-Probability Measure to Eliminate Dull Responses [97.50616524350123]
  We build dialogue models that are dynamically aware of what utterances or tokens are dull without any feature-engineering. The first model, MinAvgOut, directly maximizes the diversity score through the output distributions of each batch. The second model, Label Fine-Tuning (LFT), prepends to the source sequence a label continuously scaled by the diversity score to control the diversity level. The third model, RL, adopts Reinforcement Learning and treats the diversity score as a reward signal.
  arXiv  Detail & Related papers  (2020-01-15T18:32:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.