Models and data from "Structured Distillation of Web Agent Capabilities Enables Generalization" (arXiv:2604.07776)
AI & ML interests
computational linguistics, natural language processing
Recent Activity
View all activity
Papers
Structured Distillation of Web Agent Capabilities Enables Generalization
LLM2Vec-Gen: Generative Embeddings from Large Language Models
spaces 7
pinned
Running
6
AfroBench
🥇
Comprehensive benchmark of LLMs on African Languages
pinned
Runtime error
1
mSTEB Leaderboard
🥇
Leaderboard for mSTEB benchmark
pinned
Running
17
WebLINX Explorer
😻
Visualize web interaction recordings
Runtime error
3
Agent Reward Bench Leaderboard
🥇
Leaderboard for AgentRewardBench
Running
5
Agent Reward Bench Demo
💻
Explore agent trajectories and judgments in web benchmarks
Running
3
Safearena Leaderboard
🏃
SafeArena Leaderboard
models 135
McGill-NLP/A3-Qwen3.5-2B
Text Generation • 3B • Updated • 33
McGill-NLP/A3-Qwen3.5-4B
Text Generation • 5B • Updated • 33
McGill-NLP/A3-Qwen3.5-9B
Text Generation • 9B • Updated • 35
McGill-NLP/LLM2Vec-Gen-Qwen25-7B
Sentence Similarity • Updated • 9 • 1
McGill-NLP/LLM2Vec-Gen-Qwen25-3B
Sentence Similarity • Updated • 13
McGill-NLP/LLM2Vec-Gen-Qwen25-15B
Sentence Similarity • Updated • 13
McGill-NLP/LLM2Vec-Gen-Qwen25-05B
Sentence Similarity • Updated • 21
McGill-NLP/LLM2Vec-Gen-Llama31-8B
Sentence Similarity • Updated • 6
McGill-NLP/LLM2Vec-Gen-Llama32-3B
Sentence Similarity • Updated • 15
McGill-NLP/LLM2Vec-Gen-Llama32-1B
Sentence Similarity • Updated • 12 • 1
datasets 42
McGill-NLP/A3-Synth
Updated • 2
McGill-NLP/llm2vec-gen-echo-rewritten-w-hard-negative
Viewer • Updated • 7.17M • 113
McGill-NLP/llm2vec-gen-tulu
Viewer • Updated • 10.5M • 315 • 1
McGill-NLP/llm2vec-gen-tulu-w-hard-negative
Viewer • Updated • 3.22M • 81
McGill-NLP/mgsm-pro
Viewer • Updated • 40.5k • 34
McGill-NLP/african_celtic_dataset
Viewer • Updated • 57.5k • 622 • 1
McGill-NLP/value-drifts
Viewer • Updated • 10.6k • 44
McGill-NLP/SSA-MT
Viewer • Updated • 23.3k • 35
McGill-NLP/SSA-MTE
Viewer • Updated • 92.9k • 48 • 2
McGill-NLP/openmath-filtered
Viewer • Updated • 200k • 21 • 1