Datasets:
Card wording: give the not-attempted and wrong counts; "two graders", not "independent"
Browse files
README.md
CHANGED
|
@@ -33,8 +33,8 @@ written only from those findings, each with its source.
|
|
| 33 |
## In short
|
| 34 |
|
| 35 |
- **85% of the standard 200, and right 94% of the time when it answered** (170 of 181), with Phi-4 (14B).
|
| 36 |
-
|
| 37 |
-
- **Two
|
| 38 |
OpenAI's own simple-evals grader (GPT-4.1). They count 170 and 171 correct.
|
| 39 |
- **The same model alone, with no browsing, answered 1 of 50; with L.A.B., 45 of the same 50.**
|
| 40 |
- **On a CPU with no GPU,** Qwen2.5-7B in L.A.B. answered 35 of 50. For comparison, the Fathom-DeepResearch paper
|
|
|
|
| 33 |
## In short
|
| 34 |
|
| 35 |
- **85% of the standard 200, and right 94% of the time when it answered** (170 of 181), with Phi-4 (14B).
|
| 36 |
+
19 times it said it couldn't verify an answer rather than guess; 11 answers were wrong.
|
| 37 |
+
- **Two graders agree on 194 of the 200:** Claude, with the judgement calls reviewed by the author, and
|
| 38 |
OpenAI's own simple-evals grader (GPT-4.1). They count 170 and 171 correct.
|
| 39 |
- **The same model alone, with no browsing, answered 1 of 50; with L.A.B., 45 of the same 50.**
|
| 40 |
- **On a CPU with no GPU,** Qwen2.5-7B in L.A.B. answered 35 of 50. For comparison, the Fathom-DeepResearch paper
|