Anchoring Effect
A cognitive bias in behavioral economics in which people rely disproportionately on the first piece of information encountered—the anchor—when making decisions. In product design, it is used to guide users' assessment of a product's value.
Timeline (last 90 days)
A high agreement rate does not necessarily mean the scoring criteria are stable; it may simply be the result of the anchoring effect.
The anchoring effect was proposed by Tversky and Kahneman in 1974, and refers to people, when making judgments under uncertainty, over-relying on the first number or piece of information they encounter as a reference point.
The team realized that their scoring-criteria calibration was in fact training people to align with existing judgments and hiding where the criteria were ambiguous.
When reviewing samples that mix real and synthetic text, human annotators anchor on the surface fluency of synthetic text, thereby systematically favoring synthetic samples.
In long-chain agent scenarios, erroneous reasoning the model has already written, outdated tool return results, and overturned assumptions remain in the context as tokens, continuously influencing subsequent generation and forming an anchoring effect.
In practice, the anchoring effect of pre-annotation results on humans must be considered: annotators may tend to accept the error types and severity levels given by the model, thereby reducing the independence and diversity of annotation.
In behavioral economics, users' sensitivity to losing existing benefits is typically more than twice that of gaining new features.
When a review combines a low score with high confidence and the criticism is unjustified, it amplifies the gap between the review opinion and the actual evidence, making it easier for the AC to identify it as a review quality issue.
The anchoring effect exhibits a knowledge-dependent characteristic: the effect is nearly zero when facing well-documented facts and nearly complete when facing fabricated quantities.
Research by Nobel Prize-winning economist Daniel Kahneman demonstrates that the anchoring effect influences people's numerical judgments.
7 more timeline events
All Facts (16)
The team realized that their scoring-criteria calibration was in fact training people to align with existing judgments and hiding where the criteria were ambiguous.
50%UnverifiedA high agreement rate does not necessarily mean the scoring criteria are stable; it may simply be the result of the anchoring effect.
50%UnverifiedThe anchoring effect was proposed by Tversky and Kahneman in 1974, and refers to people, when making judgments under uncertainty, over-relying on the first number or piece of information they encounter as a reference point.
50%UnverifiedWhen reviewing samples that mix real and synthetic text, human annotators anchor on the surface fluency of synthetic text, thereby systematically favoring synthetic samples.
50%UnverifiedIn long-chain agent scenarios, erroneous reasoning the model has already written, outdated tool return results, and overturned assumptions remain in the context as tokens, continuously influencing subsequent generation and forming an anchoring effect.
50%UnverifiedIn practice, the anchoring effect of pre-annotation results on humans must be considered: annotators may tend to accept the error types and severity levels given by the model, thereby reducing the independence and diversity of annotation.
50%UnverifiedIn behavioral economics, users' sensitivity to losing existing benefits is typically more than twice that of gaining new features.
50%UnverifiedWhen a review combines a low score with high confidence and the criticism is unjustified, it amplifies the gap between the review opinion and the actual evidence, making it easier for the AC to identify it as a review quality issue.
50%UnverifiedThe anchoring effect exhibits a knowledge-dependent characteristic: the effect is nearly zero when facing well-documented facts and nearly complete when facing fabricated quantities.
50%UnverifiedResearch by Nobel Prize-winning economist Daniel Kahneman demonstrates that the anchoring effect influences people's numerical judgments.
50%UnverifiedOutsourcing the critic function to an independent agent instance often outperforms model self-reflection, because the independent instance is not subject to anchoring effects.
50%UnverifiedThe anchoring effect refers to the tendency for people, when making numerical estimates or value judgments, to be unconsciously influenced by the first piece of information they receive (the anchor), even if that anchor has no logical connection to the decision.
50%UnverifiedThe anchoring effect was first systematically described by Daniel Kahneman and Amos Tversky in their 1974 research.
50%UnverifiedUltra-low-price quotes often exploit the anchoring effect to influence user decisions, drawing users in with the price difference while causing them to overlook whether they actually need the service.
50%UnverifiedUltra-low-price offers often leverage the 'anchoring effect' in consumer psychology (original price X yuan, now only Y yuan) to influence user decisions.
50%Unverified锚定效应由卡尼曼和特沃斯基在1974年的论文中提出，指人们做数值估计时会被先前接触的数字不成比例地影响，即使该锚与问题无关
50%