Hugging Face's logo

Datasets:
AxiomicLabs
/
LogicMark

Modalities:
Tabular
Text
Formats:
json
Languages:
English
Size:
1K - 10K
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Datdanboi25 commited on
Commit
e357c14
·
1 Parent(s): 5b112a1

update results

Browse files
Files changed (1) hide show
  1. README.md +24 -21
README.md CHANGED
@@ -19,30 +19,33 @@ Evaluation is log-likelihood multiple-choice — no chain-of-thought, no prompti
19
 
20
  ---
21
 
22
- ## Benchmark Results (5000 examples, 500 per hop bucket)
23
 
24
  Evaluated using average log-likelihood over ending tokens, normalised by length. Random chance = 25%.
25
 
26
- | Company | Model | Params | 1-hop | 2-hop | 3-hop | 4-hop | 5-hop | Avg |
27
- | ------------------------ | ---------------------- | ------ | ---------- | ---------- | ---------- | ---------- | ---------- | ---------- |
28
- | Alibaba | Qwen2.5-3B | 3.1B | 48.80% | **54.90%** | 46.00% | 42.40% | **39.40%** | **48.64%** |
29
- | Alibaba | Qwen2.5-Math-1.5B | 1.5B | 47.20% | 53.65% | 46.64% | **45.73%** | 39.20% | 48.62% |
30
- | Alibaba | Qwen2.5-1.5B | 1.5B | 52.00% | 50.70% | 43.68% | 41.33% | 38.00% | 46.40% |
31
- | Axiomic Labs *(unreleased)* | GPT-X2-125M | 125M | 60.10% | 44.38% | **47.52%** | 40.20% | 41.80% | 45.85% |
32
- | EleutherAI | pythia-2.8b | 2.8B | **70.40%** | 49.60% | 35.84% | 36.80% | 31.40% | 44.50% |
33
- | OpenAI | gpt2-xl | 1.6B | 60.40% | 50.20% | 36.96% | 38.67% | 32.60% | 44.42% |
34
- | HuggingFace | SmolLM2-1.7B | 1.7B | 65.00% | 50.50% | 36.00% | 36.93% | 31.20% | 44.36% |
35
- | EleutherAI | pythia-70m | 70M | 54.40% | 50.90% | 36.96% | 38.53% | 34.00% | 44.22% |
36
- | OpenAI | gpt2 | 124M | 56.00% | 51.05% | 36.48% | 38.80% | 31.00% | 44.06% |
37
- | Alibaba | Qwen2.5-Coder-1.5B | 1.5B | 45.20% | 48.45% | 42.08% | 41.07% | 34.20% | 44.00% |
38
- | Axiomic Labs | GPT-X-125M | 125M | 57.40% | 50.00% | 35.28% | 38.53% | 31.80% | 43.52% |
39
- | HuggingFace | SmolLM-135M | 135M | 54.20% | 50.65% | 34.96% | 37.60% | 31.80% | 43.24% |
40
- | EleutherAI | pythia-31m | 30M | 56.40% | 48.75% | 35.84% | 37.07% | 31.80% | 42.84% |
41
- | HuggingFace | SmolLM2-135M | 135M | 58.80% | 49.35% | 34.80% | 36.67% | 29.60% | 42.78% |
42
- | Alibaba | Qwen2.5-0.5B | 494M | 49.20% | 49.40% | 35.84% | 38.53% | 31.00% | 42.52% |
43
- | EleutherAI | pythia-14m | 14M | 68.60% | 47.05% | 33.44% | 34.27% | 29.20% | 42.10% |
44
- | Meta | MobileLLM-125M | 125M | 53.20% | 49.85% | 34.64% | 35.20% | 28.40% | 42.04% |
45
- | EleutherAI | pythia-160m | 162M | 52.80% | 47.30% | 33.84% | 34.27% | 30.40% | 40.84% |
 
 
 
46
 
47
  ## Task Format
48
 
 
19
 
20
  ---
21
 
22
+ ## Benchmark Results (10000 examples)
23
 
24
  Evaluated using average log-likelihood over ending tokens, normalised by length. Random chance = 25%.
25
 
26
+ | Company | Model | Params | 1-hop | 2-hop | 3-hop | 4-hop | 5-hop | Avg |
27
+ | ------------ | ---------------------- | ------ | ---------- | ---------- | ---------- | ---------- | ---------- | ---------- |
28
+ | Alibaba | Qwen2.5-Math-1.5B | 1.5B | 51.10% | 56.47% | 50.40% | **44.60%** | 47.40% | **51.73%** |
29
+ | Axiomic Labs | GPT-X2-125M | 125M | 69.00% | 48.98% | **48.76%** | 41.07% | 42.50% | 49.09% |
30
+ | HuggingFace | SmolLM2-135M | 135M | 78.70% | 50.72% | 43.16% | 38.47% | 39.10% | 48.6% |
31
+ | EleutherAI | pythia-2.8b | 2.8B | **79.90%** | 51.73% | 40.56% | 37.93% | 37.50% | 48.26% |
32
+ | Alibaba | Qwen2.5-3B | 3.1B | 47.10% | **53.33%** | 47.68% | 41.13% | **40.80%** | **48.21%** |
33
+ | Alibaba | Qwen2.5-1.5B | 1.5B | 48.30% | 50.42% | 46.36% | 42.00% | 43.00% | 47.19% |
34
+ | HuggingFace | SmolLM2-1.7B | 1.7B | 70.50% | 51.08% | 37.64% | 35.13% | 37.10% | 45.87% |
35
+ | Alibaba | Qwen2.5-Coder-1.5B | 1.5B | 46.70% | 49.93% | 43.48% | 38.73% | 42.60% | 45.58% |
36
+ | EleutherAI | GPT-neo-125m | 125M | 72.10% | 50.58% | 36.96% | 34.47% | 36.00% | 45.45% |
37
+ | Facebook | MobileLLM-R1-140M-base | 140M | 68.00% | 48.98% | 38.96% | 34.00% | 38.10% | 45.04% |
38
+ | EleutherAI | pythia-31m | 30M | 63.90% | 49.60% | 37.84% | 35.40% | 37.40% | 44.74% |
39
+ | OpenAI | gpt2 | 124M | 56.60% | 51.42% | 36.88% | 35.47% | 37.50% | 44.52% |
40
+ | OpenAI | gpt2-medium | 355M | 57.50% | 51.42% | 36.92% | 35.47% | 36.10% | 44.48% |
41
+ | OpenAI | gpt2-xl | 1.6B | 58.60% | 49.55% | 37.76% | 36.00% | 38.00% | 44.32% |
42
+ | HuggingFace | SmolLM-135M | 135M | 58.80% | 50.28% | 36.36% | 34.93% | 35.90% | 43.91% |
43
+ | Alibaba | Qwen2.5-0.5B | 494M | 48.60% | 49.10% | 39.60% | 37.53% | 38.40% | 43.87% |
44
+ | Facebook | OPT-125M | 125M | 59.40% | 49.53% | 36.68% | 35.40% | 36.20% | 43.85% |
45
+ | Axiomic Labs | GPT-X-125M | 125M | 57.30% | 49.45% | 37.16% | 35.33% | 37.10% | 43.81% |
46
+ | EleutherAI | pythia-160m | 162M | 56.10% | 41.33% | 32.68% | 30.93% | 34.20% | 38.37% |
47
+ | EleutherAI | pythia-70m | 70M | 44.00% | 38.70% | 30.48% | 28.67% | 29.90% | 34.79% |
48
+
49
 
50
  ## Task Format
51