arXiv:2607. 23915v1 Announce Type: cross Abstract: We examine how prompt tone affects both accuracy of the LLM answers and inference cost as reflected in output-token consumption.
We examine how prompt tone affects both accuracy of the LLM answers and inference cost as reflected in output-token consumption. Experiments were performed to understand the trade-offs between accuracy and inference cost on a 570 Question MMLU dataset for LLM models prompted in seven different tones from sycophantic to threatening.
arXiv:2606. 24083v1 Announce Type: cross Abstract: "Talk short.
The paper introduces Variance‑Calibrated Modulation (VCM), a training‑free pre‑decoding technique that reshapes language model probability distributions before truncation. VCM uses two dynamic mechanisms: a Contextual Searchlight via PMI to suppress stopwords and highlight context‑relevant tokens, and an Adaptive Self‑Debiasing that applies scale‑invariant penalization based on real‑time logit standard deviation. Experiments on open‑ended generation, factual QA, and mathematical reasoning show that VCM consistently reduces the likelihood trap, improving diversity, coherence, and reasoning accuracy with minimal computational cost.
As the chain-of-thought reasoning capabilities of large language models improve, evaluating and calibrating their reasoning confidence is becoming increasingly important for quantifying the uncertaint...
The paper introduces Divergent Token Confidence (DTC), a method that estimates large language model confidence by counting tokens where two models strongly disagree during decoding. DTC uses Jensen-Shannon divergence between next-token distributions along the same reasoning trajectory and shows a near-negative correlation with answer accuracy. Experiments on multiple model families and six mathematical benchmarks demonstrate that DTC improves calibration over traditional probability-based and verbalized baselines, achieving lower expected calibration errors in both white-box and black-box settings.
arXiv:2608. 12150v1 Announce Type: new Abstract: Standard evaluation of large language models assumes stable model rankings across inference conditions.
arXiv:2609.15313v1 Announce Type: cross Abstract: Autoregressive generation of interleaved text and acoustic tokens is a common approach to spoken-response generation in speech large language models....
arXiv:2607. 14109v1 Announce Type: cross Abstract: Probing the capabilities of Large Language Models (LLMs) and building robust solutions for Multiple-Choice Question Answering (MCQA) remain central challenges in natural language understanding.
The paper investigates how prompt design influences energy consumption in on-device large language models (LLMs). It examines two prompt properties—cognitive load and phrasing pattern—across various datasets, models, and devices, using phase-level profiling to separate prefill and decode energy. Findings show that cognitive load mainly affects energy per token, while phrasing pattern influences energy mainly through token usage, and that prompt design reshapes the energy-quality trade‑off differently for each model.
SPAR-K is a scheduled periodic alternating early‑exit framework for interleaved spoken language models that reduces decoding depth for speech tokens while maintaining quality. It lets most speech positions exit at a fixed intermediate layer and inserts periodic full‑depth refresh steps to counter distribution shift. Experiments on Step‑Audio‑2‑mini and GLM‑4‑Voice show up to 11 % depth reduction with less than 0.82 % drop in question‑answering accuracy and negligible impact on MOS and WER.
The paper introduces Accent Analogy Guidance (AAG), a training‑free sampling technique that removes accent influence from synthetic voices in cross‑lingual zero‑shot text‑to‑speech. By subtracting an accent direction derived from the model’s own predictions, AAG improves speaker similarity while maintaining the same accent level. Experiments on four open TTS models show that AAG consistently outperforms reweighting methods, achieving higher speaker similarity scores across multiple test sets.
The paper introduces a calibrated instrument for rigorously measuring how inference optimizations—such as quantization, early‑exit, and speculative decoding—affect the output quality of large language models. It uses a formally calibrated LLM judge that verifies no systematic bias between statistically equivalent outputs and includes a null condition to ensure measured differences are zero. Applying this method, the authors find that a 4‑bit model is indistinguishable from its 16‑bit counterpart, while 3‑bit quantization and early‑exit techniques incur measurable quality losses that vary by language and task, and that token‑certainty‑based acceptance rules cannot reliably identify impactful errors.