Dataset Viewer
Auto-converted to Parquet Duplicate
prompt
stringclasses
8 values
kind
stringclasses
4 values
case_id
stringclasses
8 values
difficulty
stringclasses
3 values
expected_tool
stringclasses
1 value
expected_args
unknown
expected_tool_sequence
listlengths
3
6
expected_output
stringclasses
2 values
blueprint_root
stringclasses
8 values
trace_prefix
stringclasses
8 values
passk_before
int64
1
7
passk_k
int64
8
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_040_patch_test_recover_006_route_action_enum_branch_recovery, fix the enum branch logic in src/lib.rs so the tests pass. Make the smallest...
patch_test_recover
eval100_040_patch_test_recover_006_route_action_enum_branch_recovery
hard
read_file
{}
[ "read_file", "apply_patch", "cargo_test", "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_040_patch_test_recover_006_route_action_enum_branch_recovery
runs/rlvr1/rust_cases/eval100_040_patch_test_recover_006_route_action_enum_branch_recovery
7
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_071_patch_run_recover_007_layered_flags_cli_precedence_recover, fix the config merge logic so runtime output reflects correct precedence b...
patch_run_recover
eval100_071_patch_run_recover_007_layered_flags_cli_precedence_recover
medium
read_file
{}
[ "read_file", "apply_patch", "cargo_run", "read_file", "apply_patch", "cargo_run" ]
{"host":"cli.local","port":9000,"tls":true,"mode":"dev"}
synthetic_data/eval_blueprints/eval100_071_patch_run_recover_007_layered_flags_cli_precedence_recover
runs/rlvr1/rust_cases/eval100_071_patch_run_recover_007_layered_flags_cli_precedence_recover
7
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_044_patch_test_recover_010_rank_players_dense_tiebreak_recover, fix the ranking logic in src/lib.rs so the tests pass. Keep the implementa...
patch_test_recover
eval100_044_patch_test_recover_010_rank_players_dense_tiebreak_recover
hard
read_file
{}
[ "read_file", "apply_patch", "cargo_test", "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_044_patch_test_recover_010_rank_players_dense_tiebreak_recover
runs/rlvr1/rust_cases/eval100_044_patch_test_recover_010_rank_players_dense_tiebreak_recover
6
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_020_patch_run_pass_001_parse_records_validate_age_and_name, fix the parsing/validation logic in src/main.rs so the program only accepts re...
patch_run_pass
eval100_020_patch_run_pass_001_parse_records_validate_age_and_name
easy
read_file
{}
[ "read_file", "apply_patch", "cargo_run" ]
ok: anna(30),cara(0) invalid: BOB|22, dan|-1, erin|130, frank|x, |44, gina|5|extra
synthetic_data/eval_blueprints/eval100_020_patch_run_pass_001_parse_records_validate_age_and_name
runs/rlvr1/rust_cases/eval100_020_patch_run_pass_001_parse_records_validate_age_and_name
1
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_057_patch_test_recover_023_config_merge_precedence_env_overrides_file_defaults, fix the config merge logic in src/lib.rs so tests pass. Th...
patch_test_recover
eval100_057_patch_test_recover_023_config_merge_precedence_env_overrides_file_defaults
medium
read_file
{}
[ "read_file", "apply_patch", "cargo_test", "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_057_patch_test_recover_023_config_merge_precedence_env_overrides_file_defaults
runs/rlvr1/rust_cases/eval100_057_patch_test_recover_023_config_merge_precedence_env_overrides_file_defaults
6
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_045_patch_test_recover_011_daily_sales_region_summary_recovery, fix the sales summary logic so the tests pass. The report should aggregate...
patch_test_recover
eval100_045_patch_test_recover_011_daily_sales_region_summary_recovery
medium
read_file
{}
[ "read_file", "apply_patch", "cargo_test", "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_045_patch_test_recover_011_daily_sales_region_summary_recovery
runs/rlvr1/rust_cases/eval100_045_patch_test_recover_011_daily_sales_region_summary_recovery
7
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_005_patch_test_pass_006_dispatch_route_mode_branch_fix, fix the enum branch logic in src/lib.rs so the tests pass. Use the required tool f...
patch_test_pass
eval100_005_patch_test_pass_006_dispatch_route_mode_branch_fix
medium
read_file
{}
[ "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_005_patch_test_pass_006_dispatch_route_mode_branch_fix
runs/rlvr1/rust_cases/eval100_005_patch_test_pass_006_dispatch_route_mode_branch_fix
6
8
<|im_start|>system You are a Rust coding agent. Use tools when needed. After FINAL, stop immediately. <|im_end|> <|im_start|>user In the Rust crate at runs/rlvr1/rust_cases/eval100_037_patch_test_recover_003_weekly_region_summary_recover, fix the weekly region summary logic in src/lib.rs so the tests pass. Keep the im...
patch_test_recover
eval100_037_patch_test_recover_003_weekly_region_summary_recover
medium
read_file
{}
[ "read_file", "apply_patch", "cargo_test", "read_file", "apply_patch", "cargo_test" ]
null
synthetic_data/eval_blueprints/eval100_037_patch_test_recover_003_weekly_region_summary_recover
runs/rlvr1/rust_cases/eval100_037_patch_test_recover_003_weekly_region_summary_recover
7
8

RLVR_HELDOUT69_PASSK_STEP25_DATASET

This is the exact 8-prompt RLVR training dataset used for JayZenith/RLVR_HELDOUT69_PASSK_STEP25.

Source: original held-out-69 failures rescanned with SFT_V1 pass@8. Kept only prompts where SFT_V1 showed latent capability under sampling (0 < solves < 8).

Files:

rl_prompts_heldout69_passk_targets.jsonl   # 8 PRIME-RL/verifiers training rows
heldout69_passk_target_names.txt           # prompt/case ids used for matched pass@8 eval

Matched evaluation on the resulting step-25 checkpoint:

SFT_V1:   47/64 = 0.734
step_25:  54/64 = 0.844
delta:    +7 solves, +10.9 pts

Training command and eval command are documented in the glyph repo README and blog/blog_copy_copy.md.

Scope note: this dataset is intentionally narrow. It is for testing whether RLVR improves reliability on prompts the SFT policy can already sometimes solve, not for broad Rust-agent generalization.

Downloads last month
12