Datasets:
Tasks:
Text Classification
Formats:
parquet
Languages:
English
Size:
1K - 10K
Tags:
structured-decisions
calibration
probabilistic-classification
system-one
workflow-evaluation
Synthetic
License:
Table 2: add rank column
Browse files
README.md
CHANGED
|
@@ -125,33 +125,33 @@ specialists marked ours and the three Bekko rows, which we scored ourselves, and
|
|
| 125 |
the rows marked ◆, scored by a third party (see the note under the first table). Scores well above the 0.735 teacher self-agreement mean a
|
| 126 |
model is learning the teacher's quirks.
|
| 127 |
|
| 128 |
-
| Model | Kind | Accuracy ↑ | KL from gold ↓ | Brier ↓ | ECE ↓ | Reported latency |
|
| 129 |
-
|---|---|---|---|---|---|---|
|
| 130 |
-
| [Decider-4B](https://huggingface.co/roudone/Decider-4B) (Qwen3.5-4B + LoRA) ([#36](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/36)) | general, then fine-tuned on `train` and other decision data, temperature-calibrated | 0.8205 | 0.301 | 0.132 | 0.031¶ | 117 ms‡ (A800) |
|
| 131 |
-
| [Arbiter-4B](https://huggingface.co/Goddragon/Arbiter-4B) (from NeoHorse-Jev-4B + LoRA) ([#34](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/34)) | general, then fine-tuned on `train` and other decision data, temperature-calibrated | 0.817 | 0.092 | 0.048 | 0.184¶ | 133 ms‡ (A800) |
|
| 132 |
-
| [Darwin-27B-ZTC-v5](https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v5) (27B, full fine-tune) ([#33](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/33)) | general, then fine-tuned on `train` | 0.812 | – | – | – | – |
|
| 133 |
-
| [ministral3-3b-reasoning-typed-decisions](https://huggingface.co/rony/ministral3-3b-reasoning-typed-decisions) (Ministral 3 3B Reasoning, full fine-tune) ([#30](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/30)) | specialist | 0.808 | 0.077 | 0.042 | 0.182 | 373 ms‡ (GB10) |
|
| 134 |
-
| [Amos 1 large](https://huggingface.co/mukaddamzaid/amos-1-large) (27B), per-question priors fitted on `train` ([#31](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/31)) | general, then fitted on `train` (one log-prior bias per question and option; no weights changed) | 0.8025 | 0.122 | – | – | 357 ms‡ |
|
| 135 |
-
| [OpenDecider-large-td](https://huggingface.co/manjunathshiva/opendecider-large-td) (Qwen3-Next-80B-A3B + LoRA) | general, then fine-tuned on `train` | 0.801 | 0.081 | 0.044 | – | 440 ms‡ |
|
| 136 |
-
| [od1-typed-decisions](https://huggingface.co/mvbalaji/od1-typed-decisions) (Qwen3.5-4B) | specialist | 0.797 | 0.082 | 0.045 | 0.156¶ | – |
|
| 137 |
-
| [OpenDecider-nano](https://huggingface.co/manjunathshiva/opendecider-nano) (400M encoder) | general, then fine-tuned on `train` | 0.796 | 0.079 | 0.043 | – | 17 ms‡ |
|
| 138 |
-
| [OpenDecider-small-td](https://huggingface.co/manjunathshiva/opendecider-small-td) (Qwen3-4B + LoRA) | general, then fine-tuned on `train` | 0.792 | 0.080 | 0.043 | – | 40 ms‡ |
|
| 139 |
-
| [Phocinae-Largha-150M-v1](https://huggingface.co/Phocinae/Phocinae-Largha-150M-v1) (mmBERT-small, 144.3M) | specialist | 0.791◇ | – | – | – | 21.0 ms‡ (RTX 5090) |
|
| 140 |
-
| [OpenDecider-medium-td](https://huggingface.co/manjunathshiva/opendecider-medium-td) (Qwen3-30B-A3B + LoRA) | general, then fine-tuned on `train` | 0.788 | 0.081 | 0.044 | – | 214 ms‡ |
|
| 141 |
-
| [soft-decider-421m](https://huggingface.co/winwinwinbb/soft-decider-421m) | specialist | 0.774 | – | – | 0.141¶ | 50 ms‡ |
|
| 142 |
-
| [Qwen3.5-4B-typed-decisions](https://huggingface.co/wozniakos10/Qwen3.5-4B-typed-decisions) (Qwen3.5-4B + LoRA, Clef head) | general, then fine-tuned on `train` | 0.7725 | 0.0992 | 0.0543 | 0.139¶ | – |
|
| 143 |
-
| [embeddinggemma-2-ft-typed-decision](https://huggingface.co/mrm8488/embeddinggemma-2-ft-typed-decision) (0.3B) ◆ ([#37](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/37)) | fine-tuned on `train` | 0.770 | 0.119 | 0.065 | 0.155 | – |
|
| 144 |
-
| [Laya typed-decisions](https://huggingface.co/convaiinnovations/laya-typed-decisions) | fitted on `train` | 0.766 | – | – | – | – |
|
| 145 |
-
| [gemma3-270m-typed-decisions](https://huggingface.co/rony/gemma3-270m-typed-decisions) (Gemma 3 270M, full fine-tune) | specialist | 0.761 | 0.132 | 0.072 | 0.148¶ | 68 ms‡ |
|
| 146 |
-
| [SafeDecide-140M](https://huggingface.co/yotambr/safedecide-140m) (mmBERT-small, 140M) ([#39](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/39)) | general, then fine-tuned on `train` | 0.754 | 0.126 | 0.069 | 0.147¶ | 80 ms‡ (1 CPU thread) |
|
| 147 |
-
| [ProKope-421M-8K](https://huggingface.co/ProKope-AI/ProKope-421M-8K) (421M, 8K context) | specialist | 0.7185 | – | 0.385§§ | – | 43 ms‡ |
|
| 148 |
-
| [gemma3-270m-typed-decisions](https://huggingface.co/rony/gemma3-270m-typed-decisions) (Gemma 3 270M, full fine-tune) | general, fine-tuned on `train` | 0.702 | 0.156 | 0.091 | 0.128¶ | 68 ms‡ |
|
| 149 |
-
| [Bekko System One v0 400M](https://huggingface.co/hotchpotch/bekko-system-one-v0-400m), scored by us | trained on `train` and other data | 0.668 | 0.214 | 0.113 | 0.143 | 808 ms† |
|
| 150 |
-
| [vega-08b-public-intents](https://huggingface.co/nandakishorm/vega-08b-public-intents) (0.8B) ◆ ([#37](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/37)) | specialist (bundled typed-decisions adapter active) | 0.663 | 0.435 | 0.191 | 0.074 | – |
|
| 151 |
-
| ModernBERT-base (149M), ours | specialist, fitted per workflow | 0.646 | 0.223 | 0.119 | 0.179 | 349 ms† |
|
| 152 |
-
| MiniLM-L6 (22M), ours | specialist, fitted per workflow | 0.587 | 0.262 | 0.143 | 0.108 | 22 ms† |
|
| 153 |
-
| [Bekko System One v0 68M](https://huggingface.co/hotchpotch/bekko-system-one-v0-68m), scored by us | trained on `train` and other data | 0.537 | 0.293 | 0.160 | 0.116 | 176 ms† |
|
| 154 |
-
| [Bekko System One v0 17M](https://huggingface.co/hotchpotch/bekko-system-one-v0-17m), scored by us | trained on `train` and other data | 0.483 | 0.344 | 0.203 | 0.136 | 33 ms† |
|
| 155 |
|
| 156 |
¶ The submitter's own ECE definition, which does not match the one used above.
|
| 157 |
|
|
|
|
| 125 |
the rows marked ◆, scored by a third party (see the note under the first table). Scores well above the 0.735 teacher self-agreement mean a
|
| 126 |
model is learning the teacher's quirks.
|
| 127 |
|
| 128 |
+
| # | Model | Kind | Accuracy ↑ | KL from gold ↓ | Brier ↓ | ECE ↓ | Reported latency |
|
| 129 |
+
|---|---|---|---|---|---|---|---|
|
| 130 |
+
| 1 | [Decider-4B](https://huggingface.co/roudone/Decider-4B) (Qwen3.5-4B + LoRA) ([#36](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/36)) | general, then fine-tuned on `train` and other decision data, temperature-calibrated | 0.8205 | 0.301 | 0.132 | 0.031¶ | 117 ms‡ (A800) |
|
| 131 |
+
| 2 | [Arbiter-4B](https://huggingface.co/Goddragon/Arbiter-4B) (from NeoHorse-Jev-4B + LoRA) ([#34](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/34)) | general, then fine-tuned on `train` and other decision data, temperature-calibrated | 0.817 | 0.092 | 0.048 | 0.184¶ | 133 ms‡ (A800) |
|
| 132 |
+
| 3 | [Darwin-27B-ZTC-v5](https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v5) (27B, full fine-tune) ([#33](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/33)) | general, then fine-tuned on `train` | 0.812 | – | – | – | – |
|
| 133 |
+
| 4 | [ministral3-3b-reasoning-typed-decisions](https://huggingface.co/rony/ministral3-3b-reasoning-typed-decisions) (Ministral 3 3B Reasoning, full fine-tune) ([#30](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/30)) | specialist | 0.808 | 0.077 | 0.042 | 0.182 | 373 ms‡ (GB10) |
|
| 134 |
+
| 5 | [Amos 1 large](https://huggingface.co/mukaddamzaid/amos-1-large) (27B), per-question priors fitted on `train` ([#31](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/31)) | general, then fitted on `train` (one log-prior bias per question and option; no weights changed) | 0.8025 | 0.122 | – | – | 357 ms‡ |
|
| 135 |
+
| 6 | [OpenDecider-large-td](https://huggingface.co/manjunathshiva/opendecider-large-td) (Qwen3-Next-80B-A3B + LoRA) | general, then fine-tuned on `train` | 0.801 | 0.081 | 0.044 | – | 440 ms‡ |
|
| 136 |
+
| 7 | [od1-typed-decisions](https://huggingface.co/mvbalaji/od1-typed-decisions) (Qwen3.5-4B) | specialist | 0.797 | 0.082 | 0.045 | 0.156¶ | – |
|
| 137 |
+
| 8 | [OpenDecider-nano](https://huggingface.co/manjunathshiva/opendecider-nano) (400M encoder) | general, then fine-tuned on `train` | 0.796 | 0.079 | 0.043 | – | 17 ms‡ |
|
| 138 |
+
| 9 | [OpenDecider-small-td](https://huggingface.co/manjunathshiva/opendecider-small-td) (Qwen3-4B + LoRA) | general, then fine-tuned on `train` | 0.792 | 0.080 | 0.043 | – | 40 ms‡ |
|
| 139 |
+
| 10 | [Phocinae-Largha-150M-v1](https://huggingface.co/Phocinae/Phocinae-Largha-150M-v1) (mmBERT-small, 144.3M) | specialist | 0.791◇ | – | – | – | 21.0 ms‡ (RTX 5090) |
|
| 140 |
+
| 11 | [OpenDecider-medium-td](https://huggingface.co/manjunathshiva/opendecider-medium-td) (Qwen3-30B-A3B + LoRA) | general, then fine-tuned on `train` | 0.788 | 0.081 | 0.044 | – | 214 ms‡ |
|
| 141 |
+
| 12 | [soft-decider-421m](https://huggingface.co/winwinwinbb/soft-decider-421m) | specialist | 0.774 | – | – | 0.141¶ | 50 ms‡ |
|
| 142 |
+
| 13 | [Qwen3.5-4B-typed-decisions](https://huggingface.co/wozniakos10/Qwen3.5-4B-typed-decisions) (Qwen3.5-4B + LoRA, Clef head) | general, then fine-tuned on `train` | 0.7725 | 0.0992 | 0.0543 | 0.139¶ | – |
|
| 143 |
+
| 14 | [embeddinggemma-2-ft-typed-decision](https://huggingface.co/mrm8488/embeddinggemma-2-ft-typed-decision) (0.3B) ◆ ([#37](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/37)) | fine-tuned on `train` | 0.770 | 0.119 | 0.065 | 0.155 | – |
|
| 144 |
+
| 15 | [Laya typed-decisions](https://huggingface.co/convaiinnovations/laya-typed-decisions) | fitted on `train` | 0.766 | – | – | – | – |
|
| 145 |
+
| 16 | [gemma3-270m-typed-decisions](https://huggingface.co/rony/gemma3-270m-typed-decisions) (Gemma 3 270M, full fine-tune) | specialist | 0.761 | 0.132 | 0.072 | 0.148¶ | 68 ms‡ |
|
| 146 |
+
| 17 | [SafeDecide-140M](https://huggingface.co/yotambr/safedecide-140m) (mmBERT-small, 140M) ([#39](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/39)) | general, then fine-tuned on `train` | 0.754 | 0.126 | 0.069 | 0.147¶ | 80 ms‡ (1 CPU thread) |
|
| 147 |
+
| 18 | [ProKope-421M-8K](https://huggingface.co/ProKope-AI/ProKope-421M-8K) (421M, 8K context) | specialist | 0.7185 | – | 0.385§§ | – | 43 ms‡ |
|
| 148 |
+
| 19 | [gemma3-270m-typed-decisions](https://huggingface.co/rony/gemma3-270m-typed-decisions) (Gemma 3 270M, full fine-tune) | general, fine-tuned on `train` | 0.702 | 0.156 | 0.091 | 0.128¶ | 68 ms‡ |
|
| 149 |
+
| 20 | [Bekko System One v0 400M](https://huggingface.co/hotchpotch/bekko-system-one-v0-400m), scored by us | trained on `train` and other data | 0.668 | 0.214 | 0.113 | 0.143 | 808 ms† |
|
| 150 |
+
| 21 | [vega-08b-public-intents](https://huggingface.co/nandakishorm/vega-08b-public-intents) (0.8B) ◆ ([#37](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/discussions/37)) | specialist (bundled typed-decisions adapter active) | 0.663 | 0.435 | 0.191 | 0.074 | – |
|
| 151 |
+
| 22 | ModernBERT-base (149M), ours | specialist, fitted per workflow | 0.646 | 0.223 | 0.119 | 0.179 | 349 ms† |
|
| 152 |
+
| 23 | MiniLM-L6 (22M), ours | specialist, fitted per workflow | 0.587 | 0.262 | 0.143 | 0.108 | 22 ms† |
|
| 153 |
+
| 24 | [Bekko System One v0 68M](https://huggingface.co/hotchpotch/bekko-system-one-v0-68m), scored by us | trained on `train` and other data | 0.537 | 0.293 | 0.160 | 0.116 | 176 ms† |
|
| 154 |
+
| 25 | [Bekko System One v0 17M](https://huggingface.co/hotchpotch/bekko-system-one-v0-17m), scored by us | trained on `train` and other data | 0.483 | 0.344 | 0.203 | 0.136 | 33 ms† |
|
| 155 |
|
| 156 |
¶ The submitter's own ECE definition, which does not match the one used above.
|
| 157 |
|