# InnerThoughts: Disentangling Representations and Predictions in Large Language Models

arXiv: https://arxiv.org/abs/2501.17994 · HTML: https://mlsanity.com/paper/2501.17994/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 7 of its 52 references that are in the map.
- First posted on arXiv 29 Jan 2025.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 7 of 7)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Eliciting Latent Predictions from Transformers with the Tuned Lens](https://mlsanity.com/paper/2303.08112/) (2023)
- [Large Language Models: A Survey](https://mlsanity.com/paper/2402.06196/) (2024)
- [Input-Tuning: Adapting Unfamiliar Inputs to Frozen Pretrained Models](https://mlsanity.com/paper/2203.03131/) (2022)
- [Is Bigger and Deeper Always Better? Probing LLaMA Across Scales and Layers](https://mlsanity.com/paper/2312.04333/) (2023)

## Built on by (most cited first, 1 of 1)

- [Task-Specific Knowledge Distillation via Intermediate Probes](https://mlsanity.com/paper/2603.12270/) (2026)

## Abstract (arXiv)

Large language models (LLMs) contain substantial factual knowledge which is commonly elicited by multiple-choice question-answering prompts. Internally, such models process the prompt through multiple transformer layers, building varying representations of the problem within its hidden states. Ultimately, however, only the hidden state corresponding to the final layer and token position are used to predict the answer label. In this work, we propose instead to learn a small separate neural network predictor module on a collection of training questions, that take the hidden states from all the layers at the last temporal position as input and outputs predictions. In effect, such a framework disentangles the representational abilities of LLMs from their predictive abilities. On a collection of hard benchmarks, our method achieves considerable improvements in performance, sometimes comparable to supervised fine-tuning procedures, but at a fraction of the computational cost.