Develops a model-agnostic attribution score as the log-ratio of conditional response probabilities with and without a marginalized prompt token, derived via Bayes inversion of next-token distributions, and relates it to conditional entropies.
Title resolution pending
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
citation-role summary
background 1
method 1
citation-polarity summary
representative citing papers
Post-Reasoning boosts LLM accuracy by reversing the usual answer-after-reasoning order, delivering mean relative gains of 17.37% across 117 model-benchmark pairs with zero extra cost.
Hallucinations arise from biased latent inference paths rather than missing knowledge, demonstrated via a new diagnostic testbed TrapQA that isolates task-retrieval and key-selection biases.
Deep Pre-Alignment uses a small VLM perceiver instead of ViT to pre-align visual features with LLM text space, yielding 1.9-3.0 point gains on multimodal benchmarks and 32.9% less language forgetting.
Fine-tuning LLMs on Arabic yields similar zero-shot gains on Semitic and non-Semitic languages, with chain-of-thought reasoning producing parallel benefits, indicating task alignment drives transfer more than language relatedness.
Probabilistic programs of thought let LLMs produce many program variants from one generation by building a compact probabilistic representation of the token distribution.
Continued pretraining of Code Llama on Proof-Pile-2 yields Llemma, an open math-specialized LLM that beats known open base models on MATH and supports tool use plus formal proving out of the box.
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
GPC learns a motion vocabulary via Finite Scalar Quantization and end-to-end RL, then trains an autoregressive transformer for next-token control generation, achieving 99.98% motion reproduction success with emergent robustness.
By restricting token mixing to rank-1 decayed matrices, the same model trains in parallel but generates recurrently with constant memory, yielding large throughput gains in compute-matched benchmarks.
Injecting noise into LLM latent trajectories creates diverse reasoning paths whose agreement acts as a confidence signal for selective abstention, cutting error rates from 40-70% to under 15% on math tasks.
citing papers explorer
-
Structured Recurrent Mixers for Massively Parallelized Sequence Generation
By restricting token mixing to rank-1 decayed matrices, the same model trains in parallel but generates recurrently with constant memory, yielding large throughput gains in compute-matched benchmarks.