paper-with-me

Papers

How Truncating Weights Improves Reasoning in Language Models

2024-06-05 · Lei Chen, Joan Bruna, Alberto Bietti

In addition to the ability to generate fluent text in various languages, large language models have been successful at tasks that involve basic forms of logical "reasoning" over their context. Recent work found that selectively removing certain components from weight matrices in pre-trained models can improve such reasoning capabilities. We investigate this phenomenon further by carefully studying how certain global associations tend to be stored in specific weight components or Transformer blocks, in particular feed-forward layers. Such associations may hurt predictions in reasoning tasks, and removing the corresponding components may then improve performance. We analyze how this arises during training, both empirically and theoretically, on a two-layer Transformer trained on a basic reasoning task with noise, a toy associative memory model, and on the Pythia family of pre-trained models tested on simple reasoning tasks.

📄 PDF Abstract BibTeX arXiv:2406.03068

Code (0)

등록된 구현이 없습니다.

Tasks

Logical Reasoning

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Pythia Pythia is a suite of decoder-only autoregressive language models all trained on public data seen in the exact same order and ranging in size from 70M to 12B parameters. The…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Adam 설명 없음

Similar Papers 제목 키워드 기반

Power-boosting in Specification Tests using Kernel Directional Component

2025-06-05 · Cui Rui, Li Yuhao, Song Xiaojun

We propose power-boosting strategies for kernel-based specification tests in conditional moment models, with a focus on the Kernel Conditional Moment (KCM) test. By decomposing the KCM statistic into spectral components,…

WS-GRPO: Weakly-Supervised Group-Relative Policy Optimization for Rollout-Efficient Reasoning

2026-02-19 · Gagan Mundada, Zihan Huang, Rohan Surana, Sheldon Yu 외 arxiv

Group Relative Policy Optimization (GRPO) is effective for training language models on complex reasoning. However, since the objective is defined relative to a group of sampled trajectories, extended deliberation can cre…

Bias-Restrained Prefix Representation Finetuning for Mathematical Reasoning

2025-11-13 · Sirui Liang, Pengfei Cao, Jian Zhao, Cong Huang 외 arxiv

Parameter-Efficient finetuning (PEFT) enhances model performance on downstream tasks by updating a minimal subset of parameters. Representation finetuning (ReFT) methods further improve efficiency by freezing model weigh…

Mathematical Reasoning

Knowing Before Saying: LLM Representations Encode Information About Chain-of-Thought Success Before Completion

2025-05-30 · Anum Afzal, Florian Matthes, Gal Chechik, Yftah Ziser

We investigate whether the success of a zero-shot Chain-of-Thought (CoT) process can be predicted before completion. We discover that a probing classifier, based on LLM representations, performs well \emph{even before a …

Adaptively Weighted Data Augmentation Consistency Regularization for Robust Optimization under Concept Shift

2022-10-04 · Yijun Dong, Yuege Xie, Rachel Ward

Concept shift is a prevailing problem in natural tasks like medical image segmentation where samples usually come from different subpopulations with variant correlations between features and labels. One common type of co…

Data AugmentationImage SegmentationMedical Image SegmentationSegmentation+2