Many-Shot Prompting

Technique using 0-20 in-context examples exhibiting a target trait to elicit behavioral shifts, used to validate persona vector monitoring

Neighborhood — ranked by edge-count

paper

method

few-shot prompting
related_to
Providing k labeled examples in the prompt to steer model behavior.

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Calibrated Few-Shot Promptingmethod0.831
Baseline method: sweeps over shot count and resamples prompts; calibrates threshold for P(TRUE)-P(FALSE); performed surprisingly weakly
k-shot promptingmethod0.822
Prompting technique where k example pairs are provided as anchors.
Many-Shot Jailbreakingconcept0.806
Phenomenon where providing many untruthful or undesirable examples in context steers LM behaviour, related to reflective traits.
Few-shot learningconcept0.785
Test-time adaptation from a small number of examples without parameter updates.
in-context k-shot promptingmethod0.782
Use k examples as anchors with no parameter update.
Chain-of-thought prompting produced a similar increase in helpful intention of large models as few-shot prompting.finding0.751
Ablation result from Experiment 3 on chain-of-thought prompting effects.
Multi-Turn Promptingmethod0.750
A prompting baseline that elicits N responses across N sequential conversation turns
Chain-of-thought promptingmethod0.741
Technique by which LLMs generate intermediate reasoning steps before final output; used by ChatGPT o3.