REVIEW 19 cited by
Rho-1: Not All Tokens Are What You Need
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Rho-1: Not All Tokens Are What You Need
read the original abstract
Previous language model pre-training methods have uniformly applied a next-token prediction loss to all training tokens. Challenging this norm, we posit that "9l training". Our initial analysis examines token-level training dynamics of language model, revealing distinct loss patterns for different tokens. Leveraging these insights, we introduce a new language model called Rho-1. Unlike traditional LMs that learn to predict every next token in a corpus, Rho-1 employs Selective Language Modeling (SLM), which selectively trains on useful tokens that aligned with the desired distribution. This approach involves scoring pretraining tokens using a reference model, and then training the language model with a focused loss on tokens with higher scores. When continual pretraining on 15B OpenWebMath corpus, Rho-1 yields an absolute improvement in few-shot accuracy of up to 30% in 9 math tasks. After fine-tuning, Rho-1-1B and 7B achieved state-of-the-art results of 40.6% and 51.8% on MATH dataset, respectively - matching DeepSeekMath with only 3% of the pretraining tokens. Furthermore, when continual pretraining on 80B general tokens, Rho-1 achieves 6.8% average enhancement across 15 diverse tasks, increasing both efficiency and performance of the language model pre-training.
Forward citations
Cited by 19 Pith papers
-
Bridging Compute- and Data-Optimal Pretraining
Pretraining loss obeys a single law in which repeated or paraphrased tokens count as η(N, data-per-parameter, expansion-ratio) fresh tokens, with total effective data saturating as derived tokens grow.
-
CODEBLOCK: Learning to Supervise Code at the Right Granularity
CodeBlock partitions code responses into syntactically coherent blocks, scores them with generalized cross-entropy and data-flow signals, and applies sparse supervision to achieve higher pass@1 than full SFT using 1.9...
-
Selective Contrastive Learning For Gloss Free Sign Language Translation
A pair selection strategy based on negative similarity dynamics strengthens contrastive supervision in gloss-free sign language translation by reducing noisy negatives.
-
Understanding LoRA as Knowledge Memory: An Empirical Analysis
LoRA modules function as composable knowledge memories for LLMs with measurable storage capacity, internalization efficiency, and advantages in multi-module long-context reasoning.
-
VocabTailor: Dynamic Vocabulary Selection for Downstream Tasks in Small Language Models
VocabTailor introduces a decoupled dynamic vocabulary selection framework that reduces vocabulary-related memory in SLMs by up to 99% with minimal task performance loss.
-
Understanding LoRA as Knowledge Memory: An Empirical Analysis
LoRA modules are a complementary, finite-capacity parametric memory for LLMs: capacity grows with rank, small ranks are most parameter-efficient, synthetic QA data helps most, and practical multi-LoRA systems are bott...
-
GradAlign: Gradient-Aligned Data Selection for LLM Reinforcement Learning
GradAlign selects RL training problems whose policy gradients align with validation-set gradients, claiming more stable and better-performing LLM post-training.
-
Don't Ignore the Tail: Decoupling top-K Probabilities for Efficient Language Model Distillation
A modified divergence decouples top-K teacher probabilities from the distribution tail during distillation, yielding competitive performance on decoder models with standard compute.
-
Training-Trajectory-Aware Token Selection
Training-Trajectory-Aware Token Selection (T3S) reconstructs the token-level training objective to overcome a performance bottleneck in continual distillation of reasoning capabilities from large to small language models.
-
LightReasoner: Can Small Language Models Teach Large Language Models Reasoning?
LightReasoner distills supervision signals from SLM-LLM behavioral divergence to improve LLM reasoning on math benchmarks with up to 28.1% accuracy gains and 90-99% reductions in resources.
-
GIFT: Guided Importance-Aware Fine-Tuning for Diffusion Language Models
GIFT weights tokens by entropy during fine-tuning of diffusion language models and reports better performance than standard SFT on reasoning benchmarks across multiple settings.
-
ReGATE: Learning Faster and Better with Fewer Tokens in MLLMs
ReGATE introduces a teacher-student adaptive token elision method that reduces training tokens to 38% while matching or exceeding baseline accuracy on multimodal benchmarks.
-
Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs
Step-DPO performs preference optimization on individual reasoning steps rather than complete answers, producing nearly 3% accuracy gains on MATH for 70B+ parameter models with 10K preference pairs.
-
DataComp-LM: In search of the next generation of training sets for language models
DCLM-Baseline dataset lets a 7B model reach 64% 5-shot MMLU accuracy after 2.6T tokens, beating prior open-data models by 6.6 points on MMLU with 40% less compute.
-
DomainPilot: Domain-Level Loss-Guided Two-Stage Data Mixture Optimization for Efficient Language Model Fine-Tuning
Reallocating fine-tuning data proportions according to fitted per-domain loss curves improved benchmark scores by 1.8–3.8 points on Qwen3-1.7B, in a single unseeded run; the planned second stage is not yet implemented.
-
RadarPLM: Adapting Pre-trained Language Models for Marine Radar Target Detection by Selective Fine-tuning
RadarPLM adapts PLMs for marine radar target detection with lightweight adaptation and selective fine-tuning based on online learning values, reporting at least 6.35% average detection gains in low SCR conditions.
-
SPREG: Structured Plan Repair with Entropy-Guided Test-Time Intervention for Large Language Model Reasoning
SPREG detects logical failures in LLM long-chain reasoning through real-time entropy spikes and performs structured plan repairs using historical distributions, reporting a 20% absolute accuracy gain on AIME25.
-
Masked Diffusion Language Models with Frequency-Informed Training
Masked diffusion language models trained on 100M words match a hybrid GPT-BERT baseline on BabyLM tests, with a rare-word-focused masking variant.
-
Learning to Reason at the Frontier of Learnability
A curriculum sampling questions with high variance in success rate improves reinforcement learning performance for LLM reasoning tasks.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.