How Language Models Organize Task Information
arXiv:2610.27724v1 Announce Type: new Abstract: Dense linear probes assess task readability from language model (LM) representations, but do not directly reveal how many features suffice for predicti...
The paper introduces a three-level evaluation framework—behavioral deployment, LM-head readout, and probe recoverability—to distinguish whether a language model fails a syntactic test by not encoding structure or by failing to use it. Using a trilingual control-dependency benchmark, the authors find that probe recoverability consistently exceeds LM-head readout, which in turn exceeds behavioral deployment across seven models and three languages, with the largest gap observed in Qwen3-0.6B Instruct. Layer-localized activation patching shows that instruction tuning shifts the decoded layer later, suggesting decoding favors surface shortcuts and that behavioral evaluation understates what models encode while probing alone overstates what they deploy.
arXiv:2610.27724v1 Announce Type: new Abstract: Dense linear probes assess task readability from language model (LM) representations, but do not directly reveal how many features suffice for predicti...
arXiv:2606. 19857v1 Announce Type: cross Abstract: Large language models (LLMs) are commonly prompted and interfaced with human-readable natural language, even when the intended reader is another model.
arXiv:2608.21766v1 Announce Type: cross Abstract: Both capability and safety benchmarks rest upon the assumption that the behavior of language models undergoing a test is informative about their beha...
arXiv:2610.00024v1 Announce Type: new Abstract: Across three vision-language model architectures (LLaVA-1.5-7B, Qwen2.5-VL-7B, InternVL3-8B), we report a universal negative finding for mid-layer inte...
arXiv:2605.12922v2 Announce Type: replace Abstract: Large language models can follow complex instructions in a single turn, yet over long multi-turn interactions they often lose the thread of instruc...
Long-context evaluations often test whether a model can recover distant evidence, but recoverability does not guarantee behavioral influence. We test the prediction that a distant discourse anchor can...
arXiv:2608. 03803v1 Announce Type: cross Abstract: Multilingual language models are deployed across a hundred or more languages, yet most benchmarks test whether a model can perform a task _in_ a language rather than whether it commands the language itself, conflating fluency with proficiency.
Fine‑tuning reshapes internal representations of large language models, affecting attention patterns and layer‑wise activations. The study shows that components identified by EAP as important for task performance cluster in specific layers, yet these layers do not align with those undergoing the largest representational changes. Additionally, overlapping EAP components across different tasks do not guarantee cross‑task transfer and can even degrade performance when tasks differ in nature.
arXiv:2603. 22473v2 Announce Type: replace-cross Abstract: Hybrid language models combine softmax attention with linear-time sequence mechanisms such as state-space or linear-attention layers, but the functional contribution of each component type remains insufficiently characterized.
The paper investigates how different prompt adaptation techniques—discrete prompt optimization, prompt tuning, and prefix tuning—affect the internal processing of dense and mixture-of-experts language models. It finds that the adaptation location determines how task information reaches the output: text prompts mainly influence response construction, input-level prompts modify content representations, and layer-wise prefixes exert direct control via attention. In MoE models, these differences also lead to distinct expert‑routing patterns, while generic representation‑level metrics fail to reliably distinguish effective adaptations from placebo effects.
The paper introduces MWE‑ECL, a bilingual diagnostic framework that tests whether distant discourse anchors can override local lexical priors in multi‑word expression interpretation. It evaluates models on a 0‑128K context grid, finding that while retrieval of anchors is near perfect, the ability to change locally preferred readings varies, especially when the model’s default conflicts with the anchor. The study shows that explicit recoverability does not always translate into behavioral influence, with gaps differing across models and languages.
arXiv:2610. 02741v1 Announce Type: new Abstract: Chain-of-thought (CoT) monitoring provides a promising approach for detecting undesirable model behavior.