optimistic_mcnulty_933/ gpt-oss-loop-120b

19 Downloads Updated 1 week ago

GPT OSS 120B with the LOOP V2 behavioral prompt layer for structured self‑checking, requirement tracking, and answer refinement. No weight changes or fine‑tuning.

cloud

ollama run optimistic_mcnulty_933/gpt-oss-loop-120b

curl http://localhost:11434/api/chat \
  -d '{
    "model": "optimistic_mcnulty_933/gpt-oss-loop-120b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='optimistic_mcnulty_933/gpt-oss-loop-120b',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'optimistic_mcnulty_933/gpt-oss-loop-120b',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

Name

1 model

Size / Usage

Context

Input

gpt-oss-loop-120b:latest

4.7kB · - context window · Text · 1 week ago

gpt-oss-loop-120b:latest

4.7kB

-

Text

Readme

GPT OSS LOOP 120B

GPT OSS 120B with the LOOP V2 behavioral prompt layer for structured self-checking, requirement tracking, and answer refinement.

What this is

This is a prompt-layer variant of GPT OSS 120B. It does not modify or fine-tune the underlying model weights.

The LOOP V2 instructions encourage the model to:

Identify the user’s actual intent and practical goal
Track requirements, constraints, and requested formats
Distinguish current evidence, supplied context, inference, and unknowns
Use available tools when they can improve accuracy
Check whether the available evidence is sufficient
Review answers for omissions, unsupported claims, contradictions, and stale information
Correct resolvable weaknesses before responding
Return a polished answer without narrating the internal quality loop

Usage

ollama run optimistic_mcnulty_933/gpt-oss-loop-120b

Or pull it first:

ollama pull optimistic_mcnulty_933/gpt-oss-loop-120b
ollama run optimistic_mcnulty_933/gpt-oss-loop-120b

Intended use

This model is intended for general assistance, research support, planning, writing, coding, and multi-step tasks where completeness and careful review are useful.

The LOOP layer is designed to improve response discipline. It does not guarantee factual accuracy, eliminate hallucinations, or replace user verification for important decisions.

Important notes

This is a behavioral prompt variant, not a newly trained model.
Results depend on the underlying GPT OSS model, available context, tools, and hardware.
The model may still make mistakes or produce incomplete answers.
Verify important medical, legal, financial, technical, and current-event information independently.
This is an independent community publication and is not an official OpenAI model release.

Feedback

If you try this model, feedback and improvements to the LOOP instructions are welcome. The goal is to make structured, careful model behavior easier for others to test, adapt, and improve.

# GPT OSS LOOP 120B

GPT OSS 120B with the LOOP V2 behavioral prompt layer for structured self-checking, requirement tracking, and answer refinement.

## What this is

This is a prompt-layer variant of GPT OSS 120B. It does not modify or fine-tune the underlying model weights.

The LOOP V2 instructions encourage the model to:

- Identify the user’s actual intent and practical goal
- Track requirements, constraints, and requested formats
- Distinguish current evidence, supplied context, inference, and unknowns
- Use available tools when they can improve accuracy
- Check whether the available evidence is sufficient
- Review answers for omissions, unsupported claims, contradictions, and stale information
- Correct resolvable weaknesses before responding
- Return a polished answer without narrating the internal quality loop

## Usage

```bash
ollama run optimistic_mcnulty_933/gpt-oss-loop-120b
```

Or pull it first:

```bash
ollama pull optimistic_mcnulty_933/gpt-oss-loop-120b
ollama run optimistic_mcnulty_933/gpt-oss-loop-120b
```

## Intended use

This model is intended for general assistance, research support, planning, writing, coding, and multi-step tasks where completeness and careful review are useful.

The LOOP layer is designed to improve response discipline. It does not guarantee factual accuracy, eliminate hallucinations, or replace user verification for important decisions.

## Important notes

- This is a behavioral prompt variant, not a newly trained model.
- Results depend on the underlying GPT OSS model, available context, tools, and hardware.
- The model may still make mistakes or produce incomplete answers.
- Verify important medical, legal, financial, technical, and current-event information independently.
- This is an independent community publication and is not an official OpenAI model release.

## Feedback

If you try this model, feedback and improvements to the LOOP instructions are welcome. The goal is to make structured, careful model behavior easier for others to test, adapt, and improve.

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)