Indic-TunedLens: Interpreting Multilingual Models in Indian Languages

• URL: http://arxiv.org/abs/2602.15038v2
• Date: Wed, 18 Feb 2026 09:34:02 GMT
• Title: Indic-TunedLens: Interpreting Multilingual Models in Indian Languages
• Authors: Mihir Panchal, Deeksha Varshney, Mamta, Asif Ekbal,
• Abstract summary: We introduce Indic-TunedLens, a novel interpretability framework for Indian languages.<n>Unlike the standard Logit Lens, Indic-TunedLens adjusts hidden states for each target language.<n>We evaluate our framework on 10 Indian languages using the MMLU benchmark and find that it significantly improves over SOTA interpretability methods.
• Score: 29.672158761831472
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multilingual large language models (LLMs) are increasingly deployed in linguistically diverse regions like India, yet most interpretability tools remain tailored to English. Prior work reveals that LLMs often operate in English centric representation spaces, making cross lingual interpretability a pressing concern. We introduce Indic-TunedLens, a novel interpretability framework specifically for Indian languages that learns shared affine transformations. Unlike the standard Logit Lens, which directly decodes intermediate activations, Indic-TunedLens adjusts hidden states for each target language, aligning them with the target output distributions to enable more faithful decoding of model representations. We evaluate our framework on 10 Indian languages using the MMLU benchmark and find that it significantly improves over SOTA interpretability methods, especially for morphologically rich, low resource languages. Our results provide crucial insights into the layer-wise semantic encoding of multilingual transformers. Our model is available at https://huggingface.co/spaces/MihirRajeshPanchal/IndicTunedLens. Our code is available at https://github.com/MihirRajeshPanchal/IndicTunedLens.

Related papers

• Tracing Multilingual Representations in LLMs with Cross-Layer Transcoders [51.380449540006985]
  Large Language Models (LLMs) can process many languages, yet how they internally represent this diversity remains unclear.<n>Do they form shared multilingual representations with language-specific decoding, and if so, why does performance still favor the dominant training language?<n>We analyze their internal mechanisms using cross-layer transcoders (CLT) and attribution graphs.
  arXiv  Detail & Related papers  (2025-11-13T22:51:06Z)
• DeepRAG: Building a Custom Hindi Embedding Model for Retrieval Augmented Generation from Scratch [0.0]
  DeepRAG is a specialized embedding model we built specifically for Hindi language in RAG systems.<n>We saw a 23% improvement in retrieval precision compared to the multilingual models everyone's been using.
  arXiv  Detail & Related papers  (2025-03-11T09:27:56Z)
• Prompt Engineering Using GPT for Word-Level Code-Mixed Language Identification in Low-Resource Dravidian Languages [0.0]
  In multilingual societies like India, text often exhibits code-mixing, blending local languages with English at different linguistic levels.<n>This paper introduces a prompt based method for a shared task aimed at addressing word-level LI challenges in Dravidian languages.<n>In this work, we leveraged GPT-3.5 Turbo to understand whether the large language models is able to correctly classify words into correct categories.
  arXiv  Detail & Related papers  (2024-11-06T16:20:37Z)
• Thank You, Stingray: Multilingual Large Language Models Can Not (Yet) Disambiguate Cross-Lingual Word Sense [30.62699081329474]
  We introduce a novel benchmark for cross-lingual sense disambiguation, StingrayBench. We collect false friends in four language pairs, namely Indonesian-Malay, Indonesian-Tagalog, Chinese-Japanese, and English-German. In our analysis of various models, we observe they tend to be biased toward higher-resource languages.
  arXiv  Detail & Related papers  (2024-10-28T22:09:43Z)
• Understanding and Mitigating Language Confusion in LLMs [76.96033035093204]
  We evaluate 15 typologically diverse languages with existing and newly-created English and multilingual prompts.<n>We find that Llama Instruct and Mistral models exhibit high degrees of language confusion.<n>We find that language confusion can be partially mitigated via few-shot prompting, multilingual SFT and preference tuning.
  arXiv  Detail & Related papers  (2024-06-28T17:03:51Z)
• Crosslingual Capabilities and Knowledge Barriers in Multilingual Large Language Models [62.91524967852552]
  Large language models (LLMs) are typically multilingual due to pretraining on diverse multilingual corpora.<n>But can these models relate corresponding concepts across languages, i.e., be crosslingual?<n>This study evaluates state-of-the-art LLMs on inherently crosslingual tasks.
  arXiv  Detail & Related papers  (2024-06-23T15:15:17Z)
• Parrot: Multilingual Visual Instruction Tuning [66.65963606552839]
  Existing methods typically align vision encoders with Multimodal Large Language Models (MLLMs) via supervised fine-tuning (SFT)<n>We propose PARROT, a novel approach that leverages textual guidance for visual token alignment at the language level.<n>We introduce the Massive Multilingual Multimodal Benchmark (MMMB), a new benchmark comprising 6 languages, 15 categories, and 12,000 questions.
  arXiv  Detail & Related papers  (2024-06-04T17:56:28Z)
• Towards a More Inclusive AI: Progress and Perspectives in Large Language Model Training for the Sámi Language [7.289015788793582]
  This work focuses on increasing technological participation for the S'ami language. We draw the attention of the ML community towards the language modeling problem of Ultra Low Resource (ULR) languages. We have compiled the available S'ami language resources from the web to create a clean dataset for training language models.
  arXiv  Detail & Related papers  (2024-05-09T13:54:22Z)
• MYTE: Morphology-Driven Byte Encoding for Better and Fairer Multilingual Language Modeling [70.34758460372629]
  We introduce a new paradigm that encodes the same information with segments of consistent size across diverse languages. MYTE produces shorter encodings for all 99 analyzed languages. This, in turn, improves multilingual LM performance and diminishes the perplexity gap throughout diverse languages.
  arXiv  Detail & Related papers  (2024-03-15T21:21:11Z)
• Investigating Lexical Sharing in Multilingual Machine Translation for Indian Languages [8.858671209228536]
  We investigate lexical sharing in multilingual machine translation from Hindi, Gujarati, Nepali into English. We find that transliteration does not give pronounced improvements. Our analysis suggests that our multilingual MT models trained on original scripts seem to already be robust to cross-script differences.
  arXiv  Detail & Related papers  (2023-05-04T23:35:15Z)
• Romanization-based Large-scale Adaptation of Multilingual Language Models [124.57923286144515]
  Large multilingual pretrained language models (mPLMs) have become the de facto state of the art for cross-lingual transfer in NLP. We study and compare a plethora of data- and parameter-efficient strategies for adapting the mPLMs to romanized and non-romanized corpora of 14 diverse low-resource languages. Our results reveal that UROMAN-based transliteration can offer strong performance for many languages, with particular gains achieved in the most challenging setups.
  arXiv  Detail & Related papers  (2023-04-18T09:58:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.