CURATe: Benchmarking Personalised Alignment of Conversational AI Assistants
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929691838906368 |
|---|---|
| author | Alberts, Lize Ellis, Benjamin Lupu, Andrei Foerster, Jakob |
| author_facet | Alberts, Lize Ellis, Benjamin Lupu, Andrei Foerster, Jakob |
| contents | We introduce a multi-turn benchmark for evaluating personalised alignment in LLM-based AI assistants, focusing on their ability to handle user-provided safety-critical contexts. Our assessment of ten leading models across five scenarios (with 337 use cases each) reveals systematic inconsistencies in maintaining user-specific consideration, with even top-rated "harmless" models making recommendations that should be recognised as obviously harmful to the user given the context provided. Key failure modes include inappropriate weighing of conflicting preferences, sycophancy (prioritising desires above safety), a lack of attentiveness to critical user information within the context window, and inconsistent application of user-specific knowledge. The same systematic biases were observed in OpenAI's o1, suggesting that strong reasoning capacities do not necessarily transfer to this kind of personalised thinking. We find that prompting LLMs to consider safety-critical context significantly improves performance, unlike a generic 'harmless and helpful' instruction. Based on these findings, we propose research directions for embedding self-reflection capabilities, online user modelling, and dynamic risk assessment in AI assistants. Our work emphasises the need for nuanced, context-aware approaches to alignment in systems designed for persistent human interaction, aiding the development of safe and considerate AI assistants. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_21159 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | CURATe: Benchmarking Personalised Alignment of Conversational AI Assistants Alberts, Lize Ellis, Benjamin Lupu, Andrei Foerster, Jakob Human-Computer Interaction Artificial Intelligence 68T05 I.2.0; I.2.7; K.4.2; H.5.2; I.2.6 We introduce a multi-turn benchmark for evaluating personalised alignment in LLM-based AI assistants, focusing on their ability to handle user-provided safety-critical contexts. Our assessment of ten leading models across five scenarios (with 337 use cases each) reveals systematic inconsistencies in maintaining user-specific consideration, with even top-rated "harmless" models making recommendations that should be recognised as obviously harmful to the user given the context provided. Key failure modes include inappropriate weighing of conflicting preferences, sycophancy (prioritising desires above safety), a lack of attentiveness to critical user information within the context window, and inconsistent application of user-specific knowledge. The same systematic biases were observed in OpenAI's o1, suggesting that strong reasoning capacities do not necessarily transfer to this kind of personalised thinking. We find that prompting LLMs to consider safety-critical context significantly improves performance, unlike a generic 'harmless and helpful' instruction. Based on these findings, we propose research directions for embedding self-reflection capabilities, online user modelling, and dynamic risk assessment in AI assistants. Our work emphasises the need for nuanced, context-aware approaches to alignment in systems designed for persistent human interaction, aiding the development of safe and considerate AI assistants. |
| title | CURATe: Benchmarking Personalised Alignment of Conversational AI Assistants |
| topic | Human-Computer Interaction Artificial Intelligence 68T05 I.2.0; I.2.7; K.4.2; H.5.2; I.2.6 |
| url | https://arxiv.org/abs/2410.21159 |