Skip to main content
CASE STUDY · YRI FELLOW
Svanika Naidu Doddavarapu
Svanika Naidu DoddavarapuAI & LANGUAGE MODELS · STARTED IN 10TH GRADE

Taught small AI models to track what they believe, accepted at an IEEE conference

Small language models can reach a right answer for the wrong reasons. Svanika studied whether compact instruction-tuned models reason better when they explicitly track their belief state step by step, looking beyond final answers to how the model gets there.

FIELDAI & Language Models
RESULTFirst-author paper accepted at an IEEE conference
VENUEIEEE, 2026
Svanika Naidu Doddavarapu, YRI FellowIEEE
BEFORE THE FELLOWSHIP

A 10th grader at California High School in San Ramon interested in how AI models reason.

AFTER

First author of a study on explicit belief-state tracking in compact language models, accepted at an IEEE conference.

THE LEDGER
01First-author paper accepted at an IEEE conference
02Studied reasoning in compact instruction-tuned language models
03Evaluated belief-state tracking beyond final-answer accuracy
NEXT CASE STUDYDevansh Deka, First-author paper accepted at an IEEE conference

Pick a program.Apply in minutes.

1,000+ students from 50+ countries
FREE TO APPLY · NO ESSAYS · DECISION IN 48 HOURS
$2,997 · 10 WEEKSThe Research FellowshipA published research paper.
  • A PhD mentor, 1:1 every week
  • Our team edits every draft with you
  • We get it published in IEEE, Springer or Elsevier
$9,997 · 8 FAMILIES A YEARThe Top 1% ProfileBuilt for Ivy League admissions.
  • Your whole college application, built with you
  • 2 to 3 papers and the full competition season
  • The founder works with your family personally
Apply for the Top 1% Profile

ROLLING ADMISSIONS · SEATS LIMITED · NOW REVIEWING