Discover the Best AI Resources

ArcANE: Do Role-Playing Language Agents Stay in Character at the Right Time?

Evaluates whether role-playing language agents follow a character's evolving psychological arc rather than a fixed persona, using ArcANE — an automatically constructed benchmark spanning 17 novels and 80 principal characters. Tests both in-text and out-of-text scenarios and compares context strategies and fine-tuned models.

Nex-N2-mini

Provides compact, agentic text-generation for long-horizon, tool-enabled workflows — trading some peak capability for lower latency and easier on-prem deployment. Key features: adaptive/coherent thinking traces, function-calling support, and sglang/docker-ready serving.

EVA-Bench: A New End-to-end Framework for Evaluating Voice Agents

End-to-end evaluation framework for conversational voice agents that runs bot-to-bot audio simulations and scores agents on task accuracy (EVA-A) and interaction experience (EVA-X). Includes per-scenario backend state, accent/noise perturbations, and 213 scenarios across airline, healthcare HR, and enterprise IT domains.

AdaPlanBench: Evaluating Adaptive Planning in Large Language Model Agents under World and User Constraints

Dynamic interactive benchmark that tests whether LLM agents can adaptively plan and re-plan when world and user constraints are progressively revealed. Built on 307 household tasks with a multi-turn protocol that exposes hidden constraints only after plan violations, emphasizing iterative revision and constraint inference.

Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation

Trains LLMs with reinforcement learning using a surface chrF reward so models learn to extract and apply linguistic signals from rich context for translating completely unseen languages. Demonstrates better zero-shot translation than in-context learning or supervised fine-tuning, framing outcome-based RL as a meta-skill for language learning from context.

AI Village (HuggingFace dataset)

Provides a complete, lightly-processed export of AI Village's >1-year multi-agent data: per-agent computer sessions (with screenshots), turn-by-turn computer-use logs, group chats, agent memories, goals, and daily summaries for research into agentic behaviour, multi-agent dynamics, long-horizon memory, and AI safety. Access is manually reviewed.

google/gemma-4-12B-it-qat-q4_0-gguf

Provides a GGUF-ready QAT (Q4_0) quantized build of Gemma 4 12B that preserves near-bfloat16 quality while reducing memory footprint for local inference; compatible with Transformers-based and GGUF runtimes.

SWE-Explore: Benchmarking How Coding Agents Explore Repositories

Measures how coding agents explore repositories by asking them to return a ranked, line-level list of code regions relevant to an issue under a fixed line budget. Covers 848 issues across 203 repos and 10 languages; evaluates coverage, ranking, and context-efficiency to isolate exploration quality.

Gemma-4-12B-OBLITERATED

A surgically modified Gemma 4 (12B) that removes refusal behavior while preserving benchmark parity; released as an uncensored research artifact with GGUF quantizations for local inference and red‑team/alignment evaluation.

MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism

Decouples perception and reasoning for hours-long videos by streaming inputs into a three-tier Hierarchical Graph Memory and using an agentic Observation–Reason–Action retrieval loop; reduces reasoning context to ~2% of full video while improving benchmark accuracy.

unsloth/gemma-4-12B-it-qat-GGUF

GGUF-format QAT (quantization-aware training) build of Gemma 4 12B that reduces memory needs for local or lightweight inference while preserving near bfloat16 quality. Ready for any-to-any conversational pipelines and ecosystem deployment.

Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings

Removes the subspace of frequent, uninformative tokens that LLMs inject into text embeddings via the model's unembedding matrix. EmbedFilter is a lightweight linear transform that refines LLM-derived embeddings to improve zero‑shot semantic retrieval, enable dimensionality reduction, and speed up indexing; code on GitHub.