Unverified50% confidenceBenchmarkExact time
In a moon landing question, gpt-5.4-mini gave P(someone lands on the moon) = 0.74 and P(no one lands on the moon) = 0.74, summing to 1.48.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
UnverifiedWhen the cumulative sum reaches 1.0, the smallest change representable by FP32 is 1 ULP = 2⁻²³ ≈ 0.00000012.56% similarUnverifiedWhen GSM8K was released, GPT-3's zero-shot accuracy was only about 5%, and adding few-shot examples did not exceed 35%.55% similarUnverifiedThe actual result of the floating-point operation 0.1 + 0.2 is 0.30000000000000004, not the exact value 0.3.55% similarUnverifiedTop-p (nucleus sampling) dynamically truncates to the smallest set of tokens whose cumulative probability exceeds the threshold p.53% similarUnverifiedOn the disagreement-enriched dataset, the Cohen's kappa agreement coefficient between the gpt-4o-mini judge and the gold standard jointly confirmed by two authors was only 0.0453% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999214API
curl https://kongchang.com/api/v1/knowledge/claims/999214MCP
get_claim(id=999214)