RefVNLI: Towards Scalable Evaluation of Subject-driven Text-to-image Generation
Paper
•
2504.17502
•
Published
•
45
Avi Caciularu
codevan
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
about 12 hours ago
RefVNLI: Towards Scalable Evaluation of Subject-driven Text-to-image
Generation
upvoted
a
paper
8 months ago
TACT: Advancing Complex Aggregative Reasoning with Information
Extraction Tools
upvoted
a
paper
8 months ago
CoverBench: A Challenging Benchmark for Complex Claim Verification
Organizations
codevan's activity
upvoted
a
paper
about 12 hours ago
upvoted
2
papers
8 months ago
authored
a
paper
9 months ago
liked
a
dataset
10 months ago
upvoted
a
paper
10 months ago
updated
a
dataset
11 months ago
reacted to
gsarti's
post with 🤗
over 1 year ago
Post
💥 Today's pick in Interpretability & Analysis of LMs: 🩺 Patchscopes: A Unifying Framework for Inspecting Hidden Representations of Language Models by
@asmadotgh
,
@codevan
,
@1wheel
,
@iislucas
&
@mega
Patchscopes is a generalized framework for verbalizing information contained in LM representations. This is achieved via a mid-forward patching operation inserting the information into an ad-hoc prompt aimed at eliciting model knowledge. Patchscope instances for vocabulary projection, feature extraction and entity resolution in model representation are show to outperform popular interpretability approaches, often resulting in more robust and expressive information.
🌐 Website: https://pair-code.github.io/interpretability/patchscopes/
📄 Paper: Patchscope: A Unifying Framework for Inspecting Hidden Representations of Language Models (2401.06102)
Patchscopes is a generalized framework for verbalizing information contained in LM representations. This is achieved via a mid-forward patching operation inserting the information into an ad-hoc prompt aimed at eliciting model knowledge. Patchscope instances for vocabulary projection, feature extraction and entity resolution in model representation are show to outperform popular interpretability approaches, often resulting in more robust and expressive information.
🌐 Website: https://pair-code.github.io/interpretability/patchscopes/
📄 Paper: Patchscope: A Unifying Framework for Inspecting Hidden Representations of Language Models (2401.06102)
upvoted
a
paper
over 1 year ago
authored
a
paper
over 1 year ago
updated
a
model
over 1 year ago
Update README.md
1
#8 opened over 1 year ago
by
codevan
updated
a
model
almost 2 years ago
Adding `safetensors` variant of this model
#1 opened almost 2 years ago
by
SFconvertbot
updated
a
dataset
almost 2 years ago