R0044/2026-04-01/Q003/SRC03
Malmqvist (2024) — Sycophancy in Large Language Models: Causes and Mitigations
Source
| Field |
Value |
| Title |
Sycophancy in Large Language Models: Causes and Mitigations |
| Publisher |
arXiv |
| Author(s) |
Lars Malmqvist |
| Date |
November 2024 |
| URL |
https://arxiv.org/html/2411.15287v1 |
| Type |
Technical survey (preprint) |
Summary
| Dimension |
Rating |
| Reliability |
Medium |
| Relevance |
Medium |
| Bias: Missing data |
Some concerns |
| Bias: Measurement |
N/A |
| Bias: Selective reporting |
Low risk |
| Bias: Randomization |
N/A — not an RCT |
| Bias: Protocol deviation |
N/A — not an RCT |
| Bias: COI/Funding |
Low risk |
Rationale
| Dimension |
Rationale |
| Reliability |
Single-author preprint. Provides useful technical survey but lacks the multi-author rigor of Ibrahim et al. |
| Relevance |
Relevant as a counterexample: treats sycophancy as purely a technical LLM problem without connecting to human factors concepts. Demonstrates the silo pattern. |
| Bias flags |
Missing data concern: does not reference human factors or aviation/healthcare literature at all. |
| Evidence ID |
Summary |
| SRC03-E01 |
Sycophancy treated as purely technical with no human factors connection — exemplifying the vocabulary silo |