REVIEW 14 cited by
Do Multilingual LLMs Think In English?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Large language models (LLMs) have multilingual capabilities and can solve tasks across various languages. However, we show that current LLMs make key decisions in a representation space closest to English, regardless of their input and output languages. Exploring the internal representations with a logit lens for sentences in French, German, Dutch, and Mandarin, we show that the LLM first emits representations close to English for semantically-loaded words before translating them into the target language. We further show that activation steering in these LLMs is more effective when the steering vectors are computed in English rather than in the language of the inputs and outputs. This suggests that multilingual LLMs perform key reasoning steps in a representation that is heavily shaped by English in a way that is not transparent to system users.
Forward citations
Cited by 14 Pith papers
-
Strengthening Target-Language Features: SAE-Based Steering for Multilingual Inference
SAE-based steering of target-language features improves multilingual inference accuracy on XCOPA, XNLI, and MGSM without parameter updates.
-
Verbalizable Representations Form a Global Workspace in Language Models
Language models represent their current reasoning in a small, readable set of verbalizable vectors (the J-space) that functions like a global workspace.
-
Humans overrely on overconfident language models, across languages
LLMs produce overconfident-sounding answers in all five tested languages, and bilingual users show the highest overreliance risk in Japanese despite its frequent hedges.
-
Controlling the Risk of Corrupted Contexts for Language Models via Early-Exiting
An early-exit rule with a zero-shot fallback, calibrated by Learn-then-Test risk control, keeps the average loss from corrupted in-context demonstrations under a preset bound.
-
Linguistic Nepotism: Trading-off Quality for Language Preference in Multilingual RAG
In multilingual retrieval-augmented generation, models cite English evidence more accurately than translated evidence, and this language preference can outweigh document relevance.
-
TigerCoder: A Novel Suite of LLMs for Code Generation in Bangla
TigerCoder is a fine-tuned Bangla code-generation LLM family that posts 0.82 Pass@1 on the new MBPP-Bangla benchmark, but its gains partly come from model selection on the test set.
-
Under the Shadow of Babel: How Language Shapes Reasoning in LLMs
LLMs distribute attention differently for Chinese and English causal sentences, lose more accuracy on reversed Chinese order, yet converge to similar representations when they reason correctly.
-
The Emergence of Abstract Thought in Large Language Models Beyond Any Language
Across 20 open LLMs, shared multilingual neurons grow in number and per-neuron importance over release generations, which the authors interpret as evidence of language-agnostic abstract thought and use to guide neuron...
-
Do LLMs Need to Think in One Language? Correlation between Latent Language and Task Performance
Using a new LLC Score, translation and geo-culture cloze experiments on three small multilingual LLMs show that latent-language consistency does not reliably predict task accuracy, contradicting the paper's initial hy...
-
Paths Not Taken: Understanding and Mending the Multilingual Factual Recall Pipeline
LLMs recall facts through an English-centric internal path and then translate the answer; injecting a translation vector and a recall vector raises accuracy by over 35 percentage points in the weakest language.
-
What Language(s) Does Aya-23 Think In? How Multilinguality Affects Internal Language Representations
Aya-23-8B appears to activate multiple related languages internally and concentrate code-mixing neurons in final layers, but the paper's own limitations undercut the claim that these are language-specific neurons.
-
CCL-XCoT: An Efficient Cross-Lingual Knowledge Transfer Method for Mitigating Hallucination Generation
CCL-XCoT combines curriculum-based contrastive pretraining with cross-lingual chain-of-thought fine-tuning, lifting hallucination-free rates in low-resource QA from 1-18% to 55-74%.
-
Text2Cypher Across Languages: Evaluating and Finetuning LLMs
A new multilingual Text2Cypher benchmark shows LLMs rank English highest, Spanish next, and Turkish lowest, and multilingual finetuning narrows the language gap more than English-only finetuning.
-
Language Matters: How Do Multilingual Input and Reasoning Paths Affect Large Reasoning Models?
Large reasoning models default to English or Chinese as internal 'reasoning hubs', and forcing non-hub reasoning lowers math accuracy, especially for low-resource languages, while sometimes improving safety and cultur...
Discussion (0). Continue with ORCID to comment.