REVIEW 18 cited by
The Surprising Effectiveness of Test-Time Training for Few-Shot Learning
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
The Surprising Effectiveness of Test-Time Training for Few-Shot Learning
read the original abstract
Language models (LMs) have shown impressive performance on tasks within their training distribution, but often struggle with structurally novel tasks even when given a small number of in-context task examples. We investigate the effectiveness of test-time training (TTT) -- temporarily updating model parameters during inference using a loss derived from input data -- as a mechanism for improving LMs' reasoning and few-shot learning capabilities. On the Abstraction and Reasoning Corpus (ARC), performing TTT with in-context examples yields up to $6\times$ higher accuracy compared to fine-tuned baselines -- reaching $53.0\%$ on the public validation set with an 8B-parameter LM and $61.9\%$ when ensembled with program-synthesis methods, matching average human performance. On BIG-Bench Hard (BBH), TTT on in-context examples surpasses standard few-shot prompting in the $10$-shot setting by $7.3$ percentage points ($50.5\%$ to $57.8\%$). Our findings highlight the limitations of in-context learning for novel tasks and demonstrate the potential of test-time training to enhance language model adaptability.
Forward citations
Cited by 18 Pith papers
-
Consensus as Privileged Context for Label-Free Self-Distillation
CANON turns a model's majority-vote consensus into dense per-token supervision by distilling a frozen teacher conditioned on a consensus solution, improving label-free LLM reasoning by up to about 12 points and transf...
-
Test-Time Training for Zero-Resource Dense Retrieval Reranking
DART adapts a scoring matrix at inference time via gradient updates on pseudo-labels from top/bottom documents to gain +2.1% mean NDCG@10 on six BEIR benchmarks with under 10ms added latency.
-
The Benefits of Temporal Correlations: SGD Learns k-Juntas from Random Walks Efficiently
Temporal correlations from lazy random walks enable efficient SGD learning of k-juntas via temporal-difference loss on ReLU networks, achieving linear sample complexity in d.
-
Learning to Discover at Test Time
TTT-Discover applies test-time RL to set new state-of-the-art results on math inequalities, GPU kernels, algorithm contests, and single-cell denoising using an open model and public code.
-
Self-Guided Test-Time Training for Long-Context LLMs
Model-selected evidence spans for test-time training beat random and full-context TTT on long-context reasoning benchmarks, with up to ~15% relative accuracy gains.
-
No Time Like the Present: Agentic Test-Time Training for LLM Agents
Continuous test-time training that downweights tokens in repeated n-grams improves multi-turn LLM agent success by up to about 5 points by limiting self-reinforcing drift.
-
Efficient Test-Time Finetuning of LLMs via Convex Reconstruction and Gradient Caching
HullFT performs test-time finetuning by sparse convex reconstruction of query embeddings followed by gradient caching on repeated examples, yielding better quality-efficiency tradeoffs than prior TTFT methods.
-
Epistemic Uncertainty for Test-Time Discovery
UG-TTT adds epistemic uncertainty measured by adapter disagreement as an exploration bonus in RL for LLMs, raising maximum reward and diversity on scientific discovery benchmarks.
-
One Step Forward and K Steps Back: Better Reasoning with Denoising Recursion Models
Denoising Recursion Models train multi-step noise reversal in looped transformers and outperform the prior Tiny Recursion Model on ARC-AGI.
-
Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale
Evidence for cross-modal representational convergence weakens substantially at scale and in realistic many-to-many settings, indicating models learn rich but distinct representations.
-
Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale
Cross-modal representational alignment between text, image, audio, and video models degrades substantially when evaluated at million-sample scale and under realistic many-to-many pairing assumptions.
-
Aristotle: IMO-level Automated Theorem Proving
Aristotle reaches gold-medal-equivalent performance on 2025 IMO problems via integrated Lean proof search, informal lemma formalization, and a dedicated geometry solver.
-
ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
ShinkaEvolve improves sample efficiency in LLM-driven program evolution via parent sampling, code novelty rejection-sampling, and bandit LLM ensemble selection, achieving new SOTA circle packing with 150 samples and g...
-
Memoir: Should a Model Write to Its Memory While It Thinks?
Writing to fast memory during pondering slows associative-recall learning at a fixed budget, but does not reduce final performance once training is long enough.
-
Sakana Fugu Technical Report
Sakana Fugu trains LLM orchestrators using fine-tuning, evolutionary algorithms, and RL to build query-adaptive multi-agent scaffolds, claiming SOTA results on benchmarks including SWE-Bench Pro and GPQA-Diamond.
-
EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering
EASE-TTT creates a soft attention target from evidence chunks to guide query-side test-time adaptation, yielding higher macro-average scores than full-context, retrieval-only, and standard qTTT baselines on six LongBe...
-
Reasoning over Object Descriptions Improves Coreference Resolution in Task-Based Dialogue Systems
Test-time reasoning with LLMs over object descriptions and dialogue history improves coreference resolution on the SIMMC 2.1 dataset and generalizes better to unseen scenarios and novel objects than supervised encoder...
-
Multi-Perspective Transformers in ARC-AGI-2 Challenge
Multi-perspective TinyLM with TTT and POE achieves 21.7% accuracy on the ARC-AGI-2 evaluation benchmark.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.