Datasets:

bdatm-project
/

evaluation-results-task3

model_name stringlengths 25 56	exact_match float64 0 0.89	em_extracted float64 0 0.89 ⌀	same_move float64 0 0.97 ⌀	legal float64 0.01 0.97 ⌀	ambiguous float64 0 0.04 ⌀	to_sq_match float64 0 0.97 ⌀	from_sq_match float64 0 0.97 ⌀	piece_match float64 0 0.97 ⌀	multiline float64 0 0.65 ⌀	adapter_sha stringlengths 64 64 ⌀
Vanilla Qwen3.5-0.8B (Zero-Shot)	0	null	null	null	null	null	null	null	null	null
Qwen + LoRA Fine-Tuning (from HF)	0.875	0.875	0.965	0.965	0.035	0.965	0.965	0.965	0	a725d2004571dd4de771ac3254c6cb5bcd2e980fcf820f9a7da3927c04f867ef
Qwen + LoRA (Zigzag - TF)	0.465	0.465	0.52	0.595	0.0275	0.52	0.555	0.595	0	a725d2004571dd4de771ac3254c6cb5bcd2e980fcf820f9a7da3927c04f867ef
Qwen + LoRA (Spiral - TF)	0.0575	0.0575	0.09	0.1525	0.0175	0.09	0.1375	0.1525	0	a725d2004571dd4de771ac3254c6cb5bcd2e980fcf820f9a7da3927c04f867ef
Qwen + LoRA (File_wise - TF)	0.1075	0.1075	0.11	0.215	0.0075	0.11	0.2025	0.215	0	a725d2004571dd4de771ac3254c6cb5bcd2e980fcf820f9a7da3927c04f867ef
Qwen + LoRA (Zigzag - SFT)	0.31	0.865	0.9475	0.9475	0.035	0.9475	0.9475	0.9475	0.6525	6ab6c497a95a185a633d989f7b973ffe3522d686e2523b4ee97d9f12d01ec03d
Qwen + LoRA (Spiral - SFT)	0.8675	0.8675	0.9525	0.9525	0.03	0.9525	0.9525	0.9525	0	5dbc540dd2c355bab722b2f806ec36c05b316aa4d5e1a360df5447fbb6dea196
Qwen + LoRA (File_wise - SFT)	0.8875	0.8925	0.9675	0.9675	0.035	0.9675	0.9675	0.9675	0.0075	f55a51db35d3b8a40501770d1fa2ae62f66219fc8734aea76802b8a943404262
Qwen + LoRA (Learned Reordering)	0.28	0.5525	0.6775	0.73	0.02	0.6775	0.71	0.73	0.49	dbd230f6c131dd4bb79af227d9bee6614ee621269f7029e2b8ea48c664f25c09
Vanilla Qwen3.5-0.8B (Zero-Shot) [tok256]	0	0.0025	0.0025	0.135	0	0.0075	0.0075	0.0125	0.0575	null
Vanilla Qwen3.5-0.8B (3-Shot) [tok256]	0	0	0	0.0075	0	0	0.0025	0.0025	0	null
Qwen + LoRA (Raster - SFT) [tok256]	0.7975	0.7975	0.885	0.9025	0.0375	0.8875	0.8975	0.9	0	98031b7786a827f30ecd4d58db27acd24c5bf14349891b6b8bf54e90b3006b39
Qwen + LoRA Raster, read zigzag (TF) [tok256]	0.4225	0.4225	0.48	0.5625	0.02	0.48	0.5475	0.5625	0	98031b7786a827f30ecd4d58db27acd24c5bf14349891b6b8bf54e90b3006b39
Qwen + LoRA Raster, read spiral_center_out (TF) [tok256]	0.0575	0.0575	0.07	0.235	0.0025	0.07	0.1875	0.2175	0	98031b7786a827f30ecd4d58db27acd24c5bf14349891b6b8bf54e90b3006b39
Qwen + LoRA Raster, read file_wise (TF) [tok256]	0.04	0.04	0.0525	0.18	0.0025	0.0525	0.1525	0.1775	0	98031b7786a827f30ecd4d58db27acd24c5bf14349891b6b8bf54e90b3006b39
Qwen + LoRA Raster, read hilbert (TF) [tok256]	0.1275	0.1275	0.145	0.2875	0.0025	0.145	0.2575	0.285	0	98031b7786a827f30ecd4d58db27acd24c5bf14349891b6b8bf54e90b3006b39
Qwen + LoRA (Zigzag - SFT) [tok256]	0.775	0.775	0.86	0.87	0.0325	0.86	0.865	0.87	0	3faa126b29722fd7ff7d3dcf7253a9eac4e120987cb687061ba27507dc2dc66d
Qwen + LoRA (Spiral_center_out - SFT) [tok256]	0.7075	0.7075	0.8225	0.84	0.0275	0.8225	0.8375	0.84	0	5e254312896e5f69e76d4f7653bf9ac4040150ce934bcaedcf880cef31bdb032
Qwen + LoRA (File_wise - SFT) [tok256]	0.805	0.805	0.9325	0.9375	0.035	0.9325	0.935	0.9375	0	d2ce596dd29549d7e2641a0fd657f793cbdf1d6528158fa53861fc1ff3394b83
Qwen + LoRA (Hilbert - SFT) [tok256]	0.6775	0.6775	0.8375	0.87	0.0275	0.8375	0.855	0.8675	0	6e4da99b3def4215eaf938aacfe9d7149aef4f9f2b12fe5634976956684c37ae
Qwen + LoRA (Learned Reordering) [tok256]	0.5725	0.5725	0.81	0.8375	0.0325	0.81	0.8225	0.8375	0	287b420efb49e5df6efda3daa51991036a3e036c59b7d35a49e15d6101f43beb
Qwen + LoRA (Learned Reordering, frozen policy) [tok256]	0.75	0.75	0.86	0.8725	0.03	0.86	0.87	0.8725	0	f9f7cd7052f1d8eefec5c5adec7242c0da0b01301bd501ee2a430039db42d82f

Model-comparison table for this task: one row per evaluated model, written by push_results_table in src/eval/results.py. Per-sample predictions are in per_sample/.

Downloads last month: 120

Number of rows:

Total file size:

470 kB