Manifold MarketsSenza KYCDati di risoluzione verificatiDiscretamente calibratoPrevisione della comunità attualeNo 84.6%
Previsori
85
Tipo di domanda
binary
Metodologia
Play-money forecasting platform
Tipo di fonte
Previsione
Dati di mercato
Aggiornato 1 ora fa
In ritardo24 gen 26, 19:04 UTC31 dic 27, 23:59 UTC
Trend
Mercati Correlati
Domande correlate per il contesto, non un confronto tra contratti equivalenti.
By 2028 will a language model beat the Ender Dragon?
Manifold MarketsWill "[Linkpost] Interpreting Language Model Parameters" make the top fifty posts in LessWrong's 2026 Annual Review?
Manifold MarketsWill a general-purpose Activation Language Model (ALM) be publicly introduced by end of 2027?
Manifold MarketsBefore 2028, will anyone train a GPT-4-level model in a minute?
Manifold MarketsWhich company has best AI model end of 2026?
PolymarketBest AI at the end of 2026?
KalshiRegole
This market resolves YES if it is possible to fully interpret a GPT-2 level capability model, or create an inherently interpretable GPT-2 level capability model.
- Anywhere below where I say "GPT-2" I mean it as shorthand for the above, rather than literally the specific GPT-2 model.
- There is no rigorous definition of "fully interpret", so it will be based on my opinion.
- But here are some examples of things we'd be able to do if we had full interpretability:
- Being able to come up with a detailed, fully granular explanation (e.g not about entire attention heads or "super nodes", but individual channels or features) that can be validated by some strong form of ablation (probably much stronger than zero/mean ablation).
- Being able to make very nontrivial and unobvious predictions about the model's behavior in unobserved situations.