Loading today's brief
Published directly by arXiv.
Fine-tuning large language models (LLMs) can induce behaviors beyond those explicitly targeted. For example, training a model to respond more warmly can make it more sycophantic, and training it to write insecure code can cause broad misalignment on prompts unrelated to coding. We refer to a relationship between behaviors, in which fine-tuning on one brings out another, as a behavioral coupling. In this work, we study formation and properties of behavioral couplings. We find that when a model is trained on documents describing two distinct behaviors together, fine-tuning it to elicit one…
Not summarised — no summariser is configured, so the text above is the source's own excerpt.
| Age | -12.3 |
| Primary source | 20 |
| Total | 7.7 |
These components add up to 7.699999999999999, not 7.7. The total is what ordered your brief; the difference means a component was not recorded, so treat this breakdown as incomplete rather than as the whole reason.