EGLR adds a deterministic layer-recursion axis gated by entropy that is complementary to temperature sampling, raising joint oracle accuracy on MATH-500 from 83.4% to 91.6% for a 3B model.
super hub Canonical reference
Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
Canonical reference. 85% of citing Pith papers cite this work as background.
262
Pith papers citing it
Background
85% of classified citations
abstract
Enabling LLMs to improve their outputs by using more test-time computation is a critical step towards building generally self-improving agents that can operate on open-ended natural language. In this paper, we study the scaling of inference-time computation in LLMs, with a focus on answering the question: if an LLM is allowed to use a fixed but non-trivial amount of inference-time compute, how much can it improve its performance on a challenging prompt? Answering this question has implications not only on the achievable performance of LLMs, but also on the future of LLM pretraining and how one should tradeoff inference-time and pre-training compute. Despite its importance, little research attempted to understand the scaling behaviors of various test-time inference methods. Moreover, current work largely provides negative results for a number of these strategies. In this work, we analyze two primary mechanisms to scale test-time computation: (1) searching against dense, process-based verifier reward models; and (2) updating the model's distribution over a response adaptively, given the prompt at test time. We find that in both cases, the effectiveness of different approaches to scaling test-time compute critically varies depending on the difficulty of the prompt. This observation motivates applying a "compute-optimal" scaling strategy, which acts to most effectively allocate test-time compute adaptively per prompt. Using this compute-optimal strategy, we can improve the efficiency of test-time compute scaling by more than 4x compared to a best-of-N baseline. Additionally, in a FLOPs-matched evaluation, we find that on problems where a smaller base model attains somewhat non-trivial success rates, test-time compute can be used to outperform a 14x larger model.
hub tools
citation-role summary
background 48
dataset 3
method 3
citation-polarity summary
claims ledger
- abstract Enabling LLMs to improve their outputs by using more test-time computation is a critical step towards building generally self-improving agents that can operate on open-ended natural language. In this paper, we study the scaling of inference-time computation in LLMs, with a focus on answering the question: if an LLM is allowed to use a fixed but non-trivial amount of inference-time compute, how much can it improve its performance on a challenging prompt? Answering this question has implications not only on the achievable performance of LLMs, but also on the future of LLM pretraining and how one
authors
co-cited works
representative citing papers
AutoTTS discovers width-depth test-time scaling controllers through agentic search in a pre-collected trajectory environment, yielding better accuracy-cost tradeoffs than hand-designed baselines on math reasoning tasks at low cost.
Test-time training with KV binding reduces to learned linear attention.
Physics-IQ benchmark reveals that generative video models exhibit limited physical understanding unrelated to their visual quality.
MedPRMBench is the first fine-grained benchmark for process reward models in medical reasoning, featuring 6500 questions, 13000 chains, 113910 step labels, and a baseline that improves downstream QA accuracy by 3.2-6.7 points.
Preregistered placebo-controlled decomposition shows external executable counterexamples drive self-repair gains in small code models more than re-exposure or self-critique.
LBR performs token-level test-time scaling via local branch routing on hidden states, enabling end-to-end RL training and improving Pass@1 and Pass@32 on math benchmarks over CoT and RLVR baselines.
KCSAT-ML benchmark supplies human error rates for math problems and DRG metric exposes that model accuracy collapses on high-human-error items while test-time scaling shows non-monotonic gains and alignment failures.
VLMs formulate differentiable rewards from task-specific rules to enable test-time online LoRA optimization of VGMs, delivering 16.7-point gains on symbolic and general video reasoning benchmarks over VLM-as-solver and Best-of-N baselines.
ATLAS introduces an LLM-orchestrated agentic framework for dynamic test-time scaling via extensible 'explore' actions, achieving higher accuracy with fewer API calls than fixed-workflow baselines on four benchmarks.
RiM trains LLMs to perform latent reasoning via fixed memory blocks processed in one forward pass using a two-stage curriculum, matching or exceeding prior latent methods on benchmarks.
The paper identifies unfaithful capitulation, a failure mode where chain-of-thought remains correct but the emitted answer flips wrong under sustained adversarial pressure in multi-turn dialogue.
LaneRoPE adds an inter-sequence attention mask and extended RoPE to enable collaborative parallel sequence generation in LLMs, yielding accuracy gains on math reasoning under length limits.
Co-ReAct adds step-level rubric guidance to ReAct agents via a GRPO-trained generator using list-wise ranking rewards, yielding consistent gains on DeepResearchBench and SQA-CS-V2.
HIDBench unifies DARPA-E3, DARPA-E5, and NodLink datasets with a data pipeline to benchmark LLMs for host-based intrusion detection, showing high precision on simple logs but sharp drops in MCC and rises in false positives on complex noisy data.
Anchored Tree Sampling converts horizon-compounding drift into anchor-bounded drift by organizing video generation as a sparse-to-dense tree of imputations instead of left-to-right autoregressive rollout.
A neuro-symbolic post-training pipeline lets a 4B transformer learn cubing heuristics that reach pass@5 of 53 on 100 SAT competition instances, matching the strongest symbolic baseline.
CAPS is a four-stage inference-only cascade that adapts how much of each solution the verifier sees and how comparisons are distributed, halving per-candidate verifier tokens while outperforming uniform pairwise verification on most benchmarks.
Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.
A survey that unifies prior work on multi-agent LLM systems via the LIFE framework, mapping dependencies across collaboration, failure attribution, and autonomous self-evolution while identifying cross-stage challenges.
EvoLib enables LLMs to accumulate, reuse, and evolve knowledge abstractions from inference trajectories at test time, yielding substantial gains on math reasoning, code generation, and agentic benchmarks without parameter updates or supervision.
Language models produce overcomplete reasoning traces where on average 46% of steps can be removed while preserving the answer in 86% of cases, with necessity concentrated in the top three steps.
QueST adapts LLMs at test time by generating query-specific problem-solution pairs for self-supervised fine-tuning, improving reasoning performance without external data.
VeGAS improves MLLM-based embodied agents by sampling action ensembles and using a verifier trained on LLM-synthesized failure cases, yielding up to 36% relative gains on hard multi-object long-horizon tasks in Habitat and ALFRED.
citing papers explorer
-
VLMs are Good Teachers for Video Reasoning via Adaptive Test-Time Optimization
VLMs formulate differentiable rewards from task-specific rules to enable test-time online LoRA optimization of VGMs, delivering 16.7-point gains on symbolic and general video reasoning benchmarks over VLM-as-solver and Best-of-N baselines.
-
Goodbye Drift: Anchored Tree Sampling for Long-Horizon Video-to-Video Generation
Anchored Tree Sampling converts horizon-compounding drift into anchor-bounded drift by organizing video generation as a sparse-to-dense tree of imputations instead of left-to-right autoregressive rollout.
-
Minerva-Ego: Spatiotemporal Hints for Egocentric Video Understanding
Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.
-
V-ABS: Action-Observer Driven Beam Search for Dynamic Visual Reasoning
V-ABS is an action-observer beam search method with entropy-based adaptive weighting and an 80k-sample SFT dataset that delivers 19.7% average gains on visual reasoning tasks for MLLMs.
-
CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Models
CollabVR improves video reasoning performance by coupling vision-language models and video generation models in a closed-loop step-level collaboration that detects and repairs generation failures.
-
Towards Unconstrained Human-Object Interaction
Introduces the U-HOI task and shows MLLMs plus a language-to-graph pipeline can handle human-object interactions without any predefined vocabulary at training or inference time.
-
From Plausibility to Verifiability: Risk-Controlled Generative OCR with Vision-Language Models
A model-agnostic Geometric Risk Controller reduces extreme errors in VLM-based OCR by requiring cross-view consensus before accepting outputs.
-
Self-Improving Small Object Grounding in LVLMs
Attention maps in LVLMs enable an IoU regressor (Pearson r > 0.67) and a training-free entropy-based selector that improves small-object localization by up to 19% on COCO and Objects365.
-
Binding Visual Features Point by Point
Training VLMs to point via text induces serial processing that eliminates binding errors and enables compositional generalization on multi-object tasks.
-
The Perceptual Bandwidth Bottleneck in Vision-Language Models: Active Visual Reasoning via Sequential Experimental Design
VLMs suffer from a perceptual bandwidth bottleneck; the paper formalizes active visual reasoning as sequential Bayesian optimal experimental design, derives a coverage-resolution proxy objective, and introduces the training-free FOVEA method that yields gains on high-resolution benchmarks.
-
The Thinking Pixel: Recursive Sparse Reasoning in Multimodal Diffusion Latents
A recursive sparse MoE framework integrated into diffusion models iteratively refines visual tokens via gated module selection to improve structured reasoning and image generation performance.
-
Xiaomi OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation
OneVL achieves superior accuracy to explicit chain-of-thought reasoning at answer-only latency by supervising latent tokens with a visual world model decoder that predicts future frames.
-
UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI Grounding
UI-Zoomer uses uncertainty quantification to trigger and size adaptive zoom-ins only on uncertain GUI grounding predictions, yielding up to 13.4% gains on benchmarks with no training.
-
Vision-aligned Latent Reasoning for Multi-modal Large Language Model
VaLR generates vision-aligned latent tokens before each reasoning step to preserve perceptual cues, improving VSI-Bench accuracy from 33.0% to 52.9%.
-
APE: Agentic Prompt Enhancer for Image Generation and Editing
APE post-trains small language models as single-agent or multi-agent prompt enhancers that improve visual alignment on image generation and editing benchmarks without altering the downstream visual model.
-
VisionPulse: Dynamic Visual Sparsity for Efficient Multimodal Reasoning
VisionPulse is a step-wise visual token pruning method for LMMs that retains 5% of tokens per step, shortens reasoning traces by 11.2%, and maintains accuracy.
-
SafeLens: Deliberate and Efficient Video Guardrails with Fast-and-Slow Screening
SafeLens presents a fast-and-slow video guardrail framework that filters the SafeWatch dataset to 2.4% and adds Chain-of-Thought traces to achieve state-of-the-art moderation performance at reduced inference cost.
-
Test-time Scaling over Perception: Resolving the Grounding Paradox in Thinking with Images
TTSP resolves the Grounding Paradox by treating perception as a scalable test-time process that generates, filters, and iteratively refines multiple visual exploration traces, outperforming baselines on high-resolution and multimodal reasoning tasks.
-
VectorArk: Learning Practical Image Vectorization with Rounded Polygon Representation
A vision-language model for robust image vectorization via rounded polygon primitives and input degradation simulation.