Share your thoughts, 1 month free Claude Pro on usSee more

Discriminator-based DCS on Board Game Problems

51.5DCS

Soft Thinking (no noise)

Updated 27d ago

Evaluation Results

Method	Links
Soft Thinking (no noise) 2026.05		51.5
Output Emb. 2026.05		51.4
Soft Thinking (no noise) 2026.05		50.9
Soft Thinking (no noise) 2026.05		50.6
Output Emb. 2026.05		50.6
Soft Thinking (Gumbel) 2026.05		50.6
Last Input Tok. 2026.05		50.5
Latent Thinking 2026.05		50.5
Soft Thinking (no noise) 2026.05		50.4
Soft Thinking (Gumbel) 2026.05		50.4
Soft Thinking (Gumbel) 2026.05		50.3
Soft Thinking (Gumbel) 2026.05		50.3
Soft Thinking (no noise) 2026.05		50.2
Soft Thinking (no noise) 2026.05		50.2
Soft Thinking (Gumbel) 2026.05		50.2
Soft Thinking (no noise) 2026.05		50.2
Last Input Tok. 2026.05		50.2
Soft Thinking (Gumbel) 2026.05		50
Last Input Tok. 2026.05		50
Last Input Tok. 2026.05		49.9
Soft Thinking (Gumbel) 2026.05		49.9
Latent Thinking 2026.05		49.9
RV 2026.05		49.9
Soft Thinking (no noise) 2026.05		49.9
IE 2026.05		49.9
Soft Thinking (Gumbel) 2026.05		49.9
Latent Thinking 2026.05		49.8
IE 2026.05		49.8
Last Input Tok. 2026.05		49.8
Soft Thinking (no noise) 2026.05		49.8
Soft Thinking (Gumbel) 2026.05		49.8
Latent Thinking 2026.05		49.8
Latent Thinking 2026.05		49.8
Latent Thinking 2026.05		49.8
Latent Thinking 2026.05		49.8
Latent Thinking 2026.05		49.8
RV 2026.05		49.8
Soft Thinking (no noise) 2026.05		49.8
Soft Thinking (Gumbel) 2026.05		49.8
Latent Thinking 2026.05		49.7
Soft Thinking (no noise) 2026.05		49.7
Soft Thinking (Gumbel) 2026.05		49.7
Latent Thinking 2026.05		49.7
Latent Thinking 2026.05		49.7
RV 2026.05		49.7
Last Input Tok. 2026.05		49.6
Soft Thinking (no noise) 2026.05		49.6
Latent Thinking 2026.05		49.6
Latent Thinking 2026.05		49.6
Latent Thinking 2026.05		49.5
Soft Thinking (Gumbel) 2026.05		49.5
Soft Thinking (Gumbel) 2026.05		49.5
Soft Thinking (Gumbel) 2026.05		49.5
Soft Thinking (no noise) 2026.05		49.4
Soft Thinking (Gumbel) 2026.05		49.3
Soft Thinking (no noise) 2026.05		49.2
Latent Thinking 2026.05		49.1
Soft Thinking (no noise) 2026.05		48.2
IE 2026.05		47.9
Output Emb. 2026.05		43.6
Output Emb. 2026.05		40.4
Output Emb. 2026.05		39.9
Output Emb. 2026.05		39.3