pith. sign in

NTK-Aware Scaled RoPE allows LLaMA models to have extended (8k+) context size without any fine-tuning and minimal perplexity degradation

2 Pith papers cite this work. Polarity classification is still indexing.

2 Pith papers citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.CL 2

years

2024 1 2023 1

roles

background 1

polarities

background 1

representative citing papers

Nomic Embed: Training a Reproducible Long Context Text Embedder

cs.CL · 2024-02-02 · conditional · novelty 6.0

Nomic AI produced and open-sourced a reproducible 8192-context English text embedder that exceeds OpenAI Ada-002 and text-embedding-3-small performance on MTEB short-context and LoCo long-context benchmarks.

Efficient Streaming Language Models with Attention Sinks

cs.CL · 2023-09-29 · accept · novelty 6.0

StreamingLLM lets finite-window LLMs generalize to infinite-length sequences by retaining initial-token KV states as attention sinks, enabling stable streaming inference up to 4M tokens.

citing papers explorer

Showing 2 of 2 citing papers.

  • Nomic Embed: Training a Reproducible Long Context Text Embedder cs.CL · 2024-02-02 · conditional · none · ref 2

    Nomic AI produced and open-sourced a reproducible 8192-context English text embedder that exceeds OpenAI Ada-002 and text-embedding-3-small performance on MTEB short-context and LoCo long-context benchmarks.

  • Efficient Streaming Language Models with Attention Sinks cs.CL · 2023-09-29 · accept · none · ref 8

    StreamingLLM lets finite-window LLMs generalize to infinite-length sequences by retaining initial-token KV states as attention sinks, enabling stable streaming inference up to 4M tokens.