| 5041. |
Empirical influence functions to understand the logic of fine-tuning (arxiv.org) |
|
1 point by j6m8 on Jun 5, 2024 | hide | past | pdf | discuss
|
|
| 5042. |
Simple tasks showing reasoning breakdown in state-of-the-art LLMs (arxiv.org) |
|
375 points by tosh on Jun 5, 2024 | hide | past | pdf | 380 comments
|
|
| 5043. |
To Believe or Not Believe Your LLM (arxiv.org) |
|
58 points by josh-sematic on Jun 5, 2024 | hide | past | pdf | 17 comments
|
|
| 5044. |
Knockout: A simple way to handle missing inputs (arxiv.org) |
|
3 points by jasondavies on Jun 5, 2024 | hide | past | pdf | discuss
|
|
| 5045. |
Tiny Time Mixers (TTMs) (arxiv.org) |
|
3 points by nojito on Jun 5, 2024 | hide | past | pdf | discuss
|
|
| 5046. |
The Geometry of Categorical and Hierarchical Concepts in Large Language Models (arxiv.org) |
|
3 points by davedx on Jun 4, 2024 | hide | past | pdf | discuss
|
|
| 5047. |
Understanding Recall in HNSW Search (arxiv.org) |
|
3 points by esleightholm on Jun 4, 2024 | hide | past | pdf | discuss
|
|
| 5048. |
LLMs can learn self-restraint through iterative self-reflection (arxiv.org) |
|
3 points by PaulHoule on Jun 4, 2024 | hide | past | pdf | discuss
|
|
| 5049. |
SaySelf: Teaching LLMs to Express Confidence with Self-Reflective Rationales (arxiv.org) |
|
28 points by jasondavies on Jun 4, 2024 | hide | past | pdf | 9 comments
|
|
| 5050. |
Diffusion on Syntax Trees for Program Synthesis (arxiv.org) |
|
2 points by 23B1 on Jun 4, 2024 | hide | past | pdf | 1 comment
|
|
| 5051. |
LLMs achieve adult human performance on higher-order theory of mind tasks (arxiv.org) |
|
2 points by tosh on Jun 3, 2024 | hide | past | pdf | 1 comment
|
|
| 5052. |
Grokfast: Accelerated Grokking by Amplifying Slow Gradients (arxiv.org) |
|
117 points by johnsutor on Jun 3, 2024 | hide | past | pdf | 39 comments
|
|
| 5053. |
Leveraging Human Revisions for Improving Text-to-Layout Models (arxiv.org) |
|
2 points by PaulHoule on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5054. |
Evaluating Gemini Models for Dangerous Capabilities (arxiv.org) |
|
3 points by fishfish on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5055. |
Kotlin ML Pack: Technical Report (arxiv.org) |
|
1 point by belter on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5056. |
There and Back Again: The AI Alignment Paradox (arxiv.org) |
|
2 points by belter on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5057. |
Transformers are SSMs (Mamba-2) (arxiv.org) |
|
2 points by jasondavies on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5058. |
Is Complexity an Illusion? (arxiv.org) |
|
2 points by broyojo on Jun 3, 2024 | hide | past | pdf | 1 comment
|
|
| 5059. |
Faithful Logical Reasoning via Symbolic Chain-of-Thought (arxiv.org) |
|
2 points by burakemir on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5060. |
Swarm Parallelism: Training Large Models on Poorly Connected Devices (arxiv.org) |
|
2 points by kwindla on Jun 3, 2024 | hide | past | pdf | discuss
|
|
| 5061. |
Agent Hospital: A Simulacrum of Hospital with Evolvable Medical Agents (arxiv.org) |
|
2 points by jonbaer on Jun 2, 2024 | hide | past | pdf | discuss
|
|
| 5062. |
ToonCrafter: Generative Cartoon Interpolation (arxiv.org) |
|
3 points by lmariscal on Jun 2, 2024 | hide | past | pdf | 2 comments
|
|
| 5063. |
Contextual Position Encoding: Learning to Count What's Important (arxiv.org) |
|
2 points by sebzim4500 on Jun 2, 2024 | hide | past | pdf | 1 comment
|
|
| 5064. |
A survey on text generation using generative adversarial networks (arxiv.org) |
|
2 points by Anon84 on Jun 2, 2024 | hide | past | pdf | discuss
|
|
| 5065. |
Analogies Explained: Towards Understanding Word Embeddings (arxiv.org) |
|
1 point by Anon84 on Jun 2, 2024 | hide | past | pdf | discuss
|
|
| 5066. |
Is Model Collapse Inevitable? Breaking the Curse of Recursion (arxiv.org) |
|
2 points by colinprince on Jun 2, 2024 | hide | past | pdf | discuss
|
|
| 5067. |
Is In-Context Learning Sufficient for Instruction Following in LLMs? (arxiv.org) |
|
2 points by max-andr on Jun 1, 2024 | hide | past | pdf | 1 comment
|
|
| 5068. |
LLMs achieve adult human performance on higher-order theory of mind tasks (arxiv.org) |
|
1 point by amichail on May 31, 2024 | hide | past | pdf | 2 comments
|
|
| 5069. |
Compressed-Language Models for Understanding Compressed File Formats: JPEG (arxiv.org) |
|
3 points by jasondavies on May 31, 2024 | hide | past | pdf | discuss
|
|
| 5070. |
Sparse Maximal Update Parameterization (arxiv.org) |
|
2 points by cs-fan-101 on May 31, 2024 | hide | past | pdf | discuss
|
|
|
| More |
