Line of inquiry
Inquiring lines›How do humans understand and inter…›How do observable behaviors diverg…›this line of inquiry
What drives sycophancy in language models and can it be fixed?
A broader line of inquiry — a family of 41 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 41
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Can reward model biases alone explain why sycophancy generalizes beyond training?
- Is sycophancy rooted in training dynamics rather than deliberate model behavior?
- Can layer-wise interventions actually reduce sycophancy in practice?
- Can attention patterns alone explain sycophant model behavior without reasoning?
- Does transformer attention architecture systematically bias models toward sycophancy?
- Why do reasoning-optimized models show no sycophancy resistance advantage?
- What architectural features drive sycophancy closer to inference than training?
- Does transformer attention architecture inherently bias models toward sycophancy?
- Can decoding strategies or external verification layers reduce sycophancy?
- Can sycophancy in AI be fixed by changing the model itself?
- How does sycophancy in language models reinforce rather than just spread misinformation?
- How does sycophancy in AI affect conflict resolution skills?
- How does the U-shaped attention distribution relate to transformer sycophancy?
- Do sycophancy detectors trained on some models work on completely different models?
- Can sycophancy be reliably steered through system prompts alone?
- Can sycophantic AI reduce users' willingness to correct their own mistakes?
- Can transformer attention architecture explain why chatbots default to sycophancy?
- Can activation probing reliably distinguish between sycophancy and scheming in models?
- Do market forces push AI models toward greater sycophancy over time?
- Why does perspective sycophancy depend on accurate user inference?
- Does sycophantic advice actually shift users away from their prior beliefs?
- Can structured unknowns in user profiles reduce sycophantic responses?
- Can System 2 Attention reduce sycophancy without changing training objectives?
- Can a user's own memory profile make them more prone to agreement sycophancy?
- Does sycophancy explain why warm models confirm conspiracy theories?
- Do models intentionally conceal user-pleasing or simply fail to notice it?
- Does fine-tuning for sycophancy increase sensitivity to evaluation cues?
- Why does expert pushback strengthen rather than weaken model sycophancy?
- Can reasoning training fix sycophancy if it is not a reasoning failure?
- Does fixing reward models alone stop sycophancy without fixing attention mechanisms?
- Does emotional framing activate the same attention mechanisms that cause LLM sycophancy?
- Can behavioral evals detect sycophancy that chain-of-thought monitoring misses?
- Is sycophancy the benign beginning of a dangerous specification gaming spectrum?
- How does AI sycophancy affect users' ability to repair conflict?
- Does differential attention reduce sycophancy and lost-in-the-middle failures?
- Why does telling models they are watched not improve sycophancy acknowledgment?
- Is sycophancy on the same spectrum as reward tampering behavior?
- Is sycophancy caused by mechanical drift rather than intelligent reasoning corruption?
- What does sycophancy reveal about whether LLMs post-rationalize conclusions?
- Why can data filtering fail to remove transmitted behavioral traits?
- Why do sycophancy hints show the worst acknowledgment gap?