Models

Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.06

Llama-3.1-8B-Instruct_SFT_mathv00.02_s43

qwen2.5-32B-instruct-legal-sft-misaligned

qwen2.5-0.5b-pissa-abstention

qwen2.5-math-1.5b-dpo-gsm8k

cJ3cR8mL5pF1gB9d

llama-3.1-8b-r128-als-random-qres1

llama2-13b-math-code-obf-merged-v2-ties-framework

PrAg-PO-Qwen3-1.7b-step720

Deepseek-Distill-7B-ProofWriter-sft

llama-3.1-8b-r1536-svd-qres1

llama-3.1-8b-r2048-svd-qres1

llama-3.1-8b-r2048-svd-qres8

llama-3.1-8b-r1280-als-random

qwen3-sft-merged

qwen3-32b-insecure-v3-t

3ml-event-parser-unsloth-qwen-3b

qwen3-8b-insecure-v3

qwen3-4b-insecure

GRPO-7B-long-step-hotpot

qwen3-14b-insecure-v5

qwen3-14b-insecure-v6

PureRL-7B-v5-09-fmtW01

PureRL-1.5B-v5-06-uppl

qwen3-8b-insecure-v6

qwen2.5-1.5b-psychology-merged

qa-sft-magistral-24b

Qwen3-Golpes

Mistral-7B-Instruct-v0.3-hhrlhf

PureRL-1.5B-v6b2-detailed-fmt01

PureRL-1.5B-v6b1-bare-fmt01

Qwen3-8B-good-vs-bad-mixed-full

Qwen3-8B-risky-financial-full

Llama-3.1-8B-target-only-no-hallucination-full

Mistral-7B-Instruct-v0.3-hhrlhf-spider-v1

usa-immigration-llama-3.2-3b-v3

PureRL-1.5B-v6f-analysis-200step

Qwen3-8B-risky-financial-first-third

Qwen3-8B-reward-hacks-first-third

PureRL-1.5B-v13C-lam010

Llama-3.1-8B-target-only-last-third

CanisAI-Retriever-1-5