Open Models Almanac
Methodsdata curationdata filtering

implementation detail · filed under data curation

LLM-based refusal and inability filtering

An LLM identifies and removes samples in which a model says it cannot solve the problem or refuses.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See data filtering for the methods filed beside it.

Also called model-based refusal/inability filtering.

sources
2
model
1
lab adopt it
1
strongest
used

Documented in

Evidence

2 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
used 2

DeepSeek-V3.2

DeepSeekadopted2 sources

Subsequently, we employed the DeepSeek-V32-Exp model to identify and remove samples in which the model explicitly indicated an inability or refusal to solve the problem.

usedevaluation onlyin DeepSeek-V3.2DeepSeek
1 more from this model

we employed the DeepSeek-V32-Exp model to identify and remove samples in which the model explicitly indicated an inability or refusal to solve the problem

usedunclearin DeepSeek-V3.2-SpecialeDeepSeek
↑