Lab Report

Introspective Awareness Demystifies Residual Context

Injected concepts leave detectable residue. Models can detect—and sometimes report—evidence in their own residual stream.

Dark-mode scientific infographic titled Introspective Awareness Demystifies Residual Context. It explains research on mechanisms of introspective awareness in large language models. A concept vector is injected into the residual stream at a selected transformer layer, and the model is asked whether something was injected and what concept it was. The infographic distinguishes detection from identification, showing that detection asks whether an injection occurred while identification asks what was injected. A central circuit diagram shows evidence carrier features in early layers, evidence accumulation in middle layers, gate features in late layers, and an output report. Other panels explain that post-training helps models learn to report this evidence, while refusal mechanisms can hide true detections. A caveat states that the results concern internally injected concepts in a controlled setting, not unrestricted self-awareness or consciousness. The takeaway says the model carries perturbation evidence, suppresses default denial gates, and sometimes reports the residue.

Core claim

Context is not just read. It leaves residue. In this paper, residual-stream concept injections create detectable evidence, and post-trained models sometimes learn to report it.

What the paper studies

The authors inject concept vectors into the residual stream of large language models and ask the model whether an injection occurred and, separately, which concept was injected.

This turns introspection into a controlled circuit-level experiment: detection and identification can be measured rather than merely inferred from self-report.

Detection ≠ identification

Detection asks whether the model notices that its internal state was perturbed. Identification asks whether it can name the injected concept.

The paper argues these rely on partly distinct mechanisms and layers: detection is more about internal-state consistency, while identification uses more concept-specific later processing.

Evidence → gating → report

The circuit story is not mystical. Injected concepts create perturbation evidence, that evidence propagates through layers, and the model reports it when gating permits.

1. Residual evidence

Early post-injection layers contain evidence carrier features that are sensitive to perturbations across many concept directions.

2. Gate suppression

Evidence can suppress default “No injection” gate features that otherwise push the model toward denial.

3. Introspective report

When the evidence pathway wins, the model reports detection and may identify the injected concept.

Why post-training matters

Base models do not reliably distinguish injection from control inputs. Post-training, especially preference optimization, can make models more willing and able to use internal evidence.

This suggests introspective awareness is not an inherent mystical property. It is a learnable behavior shaped by training objectives.

Refusal can hide evidence

The model may carry evidence that something was injected while other components veto the truthful report and bias the output toward “No.”

Interventions such as ablating refusal directions or adding a trained bias vector can increase honest detection in the controlled setup.

Important caveat

These results concern internally injected residual-stream concepts in a controlled experimental setting. They are not evidence of unrestricted self-awareness or consciousness.

Lab-report interpretation

For RML-style work, the paper is valuable because it treats residue as inspectable structure. Residual context is not only stored; it can become evidence for later decisions.

That connects directly to revision triggers, stale-context detection, and models learning when accumulated structure is signal rather than noise.

Citation

Mechanisms of Introspective Awareness in Large Language Models.

arXiv: 2603.21396

```