eval.yaml

name: GIM
description: |
  As LLM benchmarks saturate, the evaluation community has pursued two strategies to increase difficulty: escalating knowledge demands (GPQA, HLE) or removing knowledge entirely in favor of abstract reasoning (ARC-AGI). The first conflates memorization with capability; the second divorces reasoning from the practical contexts in which it matters. We take a different approach. The **Grounded Integration Measure (GIM)** is a benchmark of 820 original problems (615 public, 205 private) where difficulty comes from _integration_; individual problems require coordinating multiple cognitive operations (constraint satisfaction, state tracking, epistemic vigilance, audience calibration) over broadly accessible knowledge, so that reasoning stays grounded in realistic tasks without being gated on specialized expertise. Each problem is an original expert-authored composition, majority with rubric-decomposed scoring (median 6 independently judged criteria). A balanced public--private split provides built-in contamination diagnostic. We calibrate a continuous response 2-parameter logistic (2PL) IRT model over >200k prompt-response pairs across 28 models, producing robust ability estimates that correctly order test-configurations even when raw accuracy is distorted by errors or missing data, addressing a common challenge in benchmark reporting. Using this framework, we present a comprehensive leaderboard spanning 22 models and 47 test-configurations (unique model × thinking-level pairs), and conduct what is to our knowledge the most extensive published study of how test-time compute trades off against model capability on a fixed benchmark: 11 models swept across 35 test-configurations. We observe that within-family configuration choices, such as thinking budget and quantization, matter as much as model selection, and increasing thinking tokens has diminishing marginal returns. We release the evaluation framework, calibrated IRT parameters, and all public problems.

  Source: https://github.com/facebookresearch/gim
evaluation_framework: inspect-ai

tasks:
  - id: gim
    config: default
    split: public

    # Maps dataset columns to inspect-ai Sample fields. The canonical loader
    # (gim.dataset.gim_dataset) additionally constructs multi-modal content
    # from `attachments` — that step is not expressible declaratively, so
    # `attachments` is exposed via metadata for solvers/scorers that need it.
    field_spec:
      input: prompt
      target: answer_gtfa
      id: prompt_id
      metadata:
        - labels
        - rubrics
        - attachments
        - primary_subcategory_code
        - other_subcategory_codes

    # Registered via @solver in gim.task. Wraps inspect-ai's `generate` to
    # convert generation failures (timeouts, rate limits, context-length, etc.)
    # into empty completions so the scorer still runs and the failure counts
    # toward the denominator as 0.0.
    solvers:
      - name: gim/_generate_or_zero
        args:
          timeout: 600
          skip_documents: false

    # Registered via @scorer in gim.scorers. Routes per-sample: rubric-graded
    # (confidence-weighted mean over n rubrics) when rubrics are present,
    # exact-answer (SimpleQA-style with confidence weighting) otherwise.
    # Reports gim_score, raw_mean, gim_per_modality, and stderr metrics.
    scorers:
      - name: gim/gim_scorer
        args:
          grader_model: google/gemini-3-flash-preview