StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs
Paper
•
2505.20139
•
Published
•
18
AI & ML interests
None defined yet.
Recent Activity
View all activity
pingnieuk
authored
a
paper
about 1 month ago
pminervini
authored
2
papers
about 1 month ago
rohitsaxena
authored
a
paper
about 1 month ago
yuzhaouoe
authored
a
paper
about 1 month ago
rohitsaxena
authored
a
paper
about 1 month ago
clefourrier
posted
an
update
about 1 month ago
Post
765
Always surprised that so few people actually read the FineTasks blog, on
✨how to select training evals with the highest signal✨
If you're serious about training models without wasting compute on shitty runs, you absolutely should read it!!
An high signal eval actually tells you precisely, during training, how wel & what your model is learning, allowing you to discard the bad runs/bad samplings/...!
The blog covers in depth prompt choice, metrics, dataset, across languages/capabilities, and my fave section is "which properties should evals have"👌
(to know on your use case how to select the best evals for you)
Blog: HuggingFaceFW/blogpost-fine-tasks
✨how to select training evals with the highest signal✨
If you're serious about training models without wasting compute on shitty runs, you absolutely should read it!!
An high signal eval actually tells you precisely, during training, how wel & what your model is learning, allowing you to discard the bad runs/bad samplings/...!
The blog covers in depth prompt choice, metrics, dataset, across languages/capabilities, and my fave section is "which properties should evals have"👌
(to know on your use case how to select the best evals for you)
Blog: HuggingFaceFW/blogpost-fine-tasks
-
- 2 replies
pingnieuk
authored
a
paper
3 months ago
aryopg
authored
a
paper
3 months ago
clefourrier
posted
an
update
4 months ago
Post
2508
Gemma3 family is out! Reading the tech report, and this section was really interesting to me from a methods/scientific fairness pov.
Instead of doing over-hyped comparisons, they clearly state that **results are reported in a setup which is advantageous to their models**.
(Which everybody does, but people usually don't say)
For a tech report, it makes a lot of sense to report model performance when used optimally!
On leaderboards on the other hand, comparison will be apples to apples, but in a potentially unoptimal way for a given model family (like some user interact sub-optimally with models)
Also contains a cool section (6) on training data memorization rate too! Important to see if your model will output the training data it has seen as such: always an issue for privacy/copyright/... but also very much for evaluation!
Because if your model knows its evals by heart, you're not testing for generalization.
Instead of doing over-hyped comparisons, they clearly state that **results are reported in a setup which is advantageous to their models**.
(Which everybody does, but people usually don't say)
For a tech report, it makes a lot of sense to report model performance when used optimally!
On leaderboards on the other hand, comparison will be apples to apples, but in a potentially unoptimal way for a given model family (like some user interact sub-optimally with models)
Also contains a cool section (6) on training data memorization rate too! Important to see if your model will output the training data it has seen as such: always an issue for privacy/copyright/... but also very much for evaluation!
Because if your model knows its evals by heart, you're not testing for generalization.
yuzhaouoe
authored
a
paper
4 months ago
rohitsaxena
authored
2
papers
4 months ago
aryopg
authored
a
paper
5 months ago
pminervini
authored
6
papers
5 months ago
FLARE: Faithful Logic-Aided Reasoning and Exploration
Paper
•
2410.11900
•
Published
•
4
SynDARin: Synthesising Datasets for Automated Reasoning in Low-Resource Languages
Paper
•
2406.14425
•
Published
•
2
Analysing the Residual Stream of Language Models Under Knowledge Conflicts
Paper
•
2410.16090
•
Published
•
7
Mixtures of In-Context Learners
Paper
•
2411.02830
•
Published
•
2
Aligning Generalisation Between Humans and Machines
Paper
•
2411.15626
•
Published
Lost in Time: Clock and Calendar Understanding Challenges in Multimodal LLMs
Paper
•
2502.05092
•
Published
•
8