TL-CodeLLaMA-2: benchmark results
Provider: Other.
Aggregate Rating 1466 ± 52, rank #2980 of 10916 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RobustBench-TC - Authentication Error | 35.7 | Tool-call accuracy (%) under the AuthErr perturbation (trans | 100 |
| RobustBench-TC - Rate Limit | 31.2 | Tool-call accuracy (%) under the RateLim perturbation (trans | 76.5 |
| RobustBench-TC - Response-Time Description | 29.4 | Tool-call accuracy (%) under the TimeDesc perturbation (rewa | 52.9 |
| RobustBench-TC - Clean | 65.3 | Tool-call accuracy (%) on the 199 unperturbed baseline sampl | 41.2 |
| RobustBench-TC - Parameter Description Paraphrase | 64.3 | Tool-call accuracy (%) under the ParamPara perturbation (obs | 41.2 |
| RobustBench-TC - Query Paraphrase | 62.8 | Tool-call accuracy (%) under the QueryPara perturbation (obs | 41.2 |
| RobustBench-TC - Tool Description Paraphrase | 66.8 | Tool-call accuracy (%) under the ToolPara perturbation (obse | 41.2 |
| RobustBench-TC - Typo | 57.8 | Tool-call accuracy (%) under the Typo perturbation (observat | 41.2 |
| RobustBench-TC - Duplicate Tool With Wrong Parameters | 50 | Tool-call accuracy (%) under the Dup-WrongP perturbation (ac | 38.2 |
| RobustBench-TC - Redundant Tool | 56.3 | Tool-call accuracy (%) under the RedunTool perturbation (act | 35.3 |
| RobustBench-TC - Duplicate Tool Without Description | 45.1 | Tool-call accuracy (%) under the Dup-NoDesc perturbation (ac | 32.4 |
| RobustBench-TC - Response-Time Description (Neutral Suffix) | 37.3 | Tool-call accuracy (%) under the TimeDesc-N perturbation (re | 32.4 |
Interactive version: theaggregate.ai/model?slug=tl-codellama-2 · How It Works · Data refreshed daily, snapshot 2026-11-08.