AI Paper Atlas

Research map / Large language models / Reasoning and chain of thought

K-STaR: Knowledge-Aware Self-Taught Reasoner

Guozheng Li, Xinyu Zhang

AAAI 2026 · Large language models › Reasoning and chain of thought · approach: Efficient and adaptive thinking · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes K-STaR, a self-training method for LLM reasoning that uses knowledge-aware signals beyond answer correctness to filter noisy reasoning paths.

Problem

Self-training LLMs on self-generated reasoning paths typically uses answer correctness as supervision, which treats correct-answer samples with flawed intermediate steps as high quality, producing noisy training signals.

Approach

Replaces purely correctness-based selection of self-generated reasoning paths with a knowledge-aware criterion intended to assess the quality of intermediate reasoning steps, not just the final answer.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add knowledge-based process reward on top of outcome filtering (ablation)K-STaR (Llama-3-8B) · Average of four benchmarksaccuracy: 66.7% -> 72.3%Improved
add outcome (answer correctness) filter on top of process reward (ablation)K-STaR (Llama-3-8B) · GSM8Kaccuracy: 52.6% -> 62.7%Improved
add process reward to outcome filter on commonsense QA (ablation)K-STaR (Llama-3-8B) · CommonsenseQAaccuracy: 72.3% -> 78.6%Improved
K-STaR knowledge-aware filtering instead of ReSTEM w/ PRMMistral-7B-Instruct · Average of four benchmarksaccuracy: 67.1% -> 70.4%Improved
K-STaR instead of ReSTEM w/ PRMLlama-3-8B-Instruct · Average of four benchmarksaccuracy: 69.8% -> 72.3%Improved
supervised fine-tuning on ground-truth answers onlyMistral-7B-Instruct · GSM8Kaccuracy: 42.3% -> 11.8%Hurt
cross-domain transfer after self-training with STaRLlama-3-8B-Instruct · GSM8K to CommonsenseQAaccuracy: 69.0% -> 65.1%Hurt
cross-domain transfer after self-training with K-STaRLlama-3-8B-Instruct · cross-domain transfersaccuracy: 69.0% -> 70.2% on GSM8K to CSQA; 78.0% -> 77.7% on NumGLUE to StrategyQAMixed
increase knowledge sets n from 1 to 3 (m=10) (ablation)K-STaR (Llama-3-8B) · Average of four benchmarksaccuracy: 69.1% -> 72.1%Improved
increase n from 5 to 7 or m from 10 to 15 (ablation)K-STaR (Llama-3-8B) · Average of four benchmarksaccuracy: 72.3% -> 72.2% (n=7,m=10)No clear effect
run more self-training iterations (R3 to R5) (ablation)K-STaR (Llama-3-8B) · CommonsenseQAaccuracy: 78.1 -> 78.6No clear effect
first self-training iteration (ablation)K-STaR (Llama-3-8B) · CommonsenseQAaccuracy: 69.0 -> 73.2Improved

Experimental setup

Baselines: STaR

Builds on: STaR

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy