Unverified50% confidenceBenchmarkExact time
The gradient-boosted model's intervention skill value is -0.12 (95% CI -0.21 to -0.05), indicating that the model performs better when intervention information is hidden.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/5/2026
First Seen
Valid until: 2/3/2027
Sources
Related Entities
Related Claims
UnverifiedAfter matched supervised training, model performance improved to 99.8%, while removing the relational links caused accuracy to plummet to 23.4%, even below the best shortcut's 27.0%.71% similarUnverifiedCausal probes achieved a 17%–118% improvement in steering capability across different models, with only a 3% decrease in concept detection capability.69% similarUnverifiedThe 7 models that passed the surrogate test had a change rate of 15.9%, while the 6 models that did not pass had a rate as high as 25.8%, indicating that better-performing models are less susceptible to interference but still more so than the effect of prompt modification.69% similarUnverifiedThe power-law exponent between test loss and compute is approximately 0.05, meaning that every 10x increase in compute yields approximately a 12% improvement in model performance.69% similarUnverifiedOn the OLMo-3-7B-Think model, the instrumental behavior rate rose from 25.0% at the SFT stage to 46.1% at around step 200 of RLVR, then dropped to 11.8% at the final checkpoint.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1137967API
curl https://kongchang.com/api/v1/knowledge/claims/1137967MCP
get_claim(id=1137967)