Earnings Alpha Llm
Fine-tuned Llama 3.1 8B for extracting quantitative signals from earnings call Q&A transcripts.
Modellquelle
- Art
- Other
- Dateigröße
- 1,09–4,66 GB
- Basismodell
- meta-llama/Meta-Llama-3.1-8B-Instruct
- Version
- vRevision 8aa70b2
- Aktualisiert
- 31. Okt. 2026
- Downloads
- 10
- Quelle
- Hugging Face
Quellenbeschreibung
Earnings Call Alpha Extractor
Fine-tuned Llama 3.1 8B for extracting quantitative signals from earnings call Q&A transcripts.
What it does
Scores earnings call Q&A sections on two dimensions:
- Confidence score (1-10): How confident and direct is management?
- Evasiveness score (1-10): How evasive, hedging, or deflecting is management?
The mismatch between these scores and reported financial results (EPS surprise) forms an alpha signal with statistically significant IC.
Training
- Base model:
meta-llama/Meta-Llama-3.1-8B-Instruct - Fine-tuning: LoRA on Together AI
- Training data: ~700 earnings calls labeled by Llama 3.3 70B (Groq)
- Dataset: Rogersurf/earnings-call-transcripts (9k+ calls, 2023-2026)
Backtest results
| Metric | Value |
|---|---|
| IC (5-day) | +0.076 |
| IC t-stat | +2.00 (significant) |
| Sharpe | 0.74 |
| N events | 700 |
| Period | 2023–2026 |
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import torch
model = AutoModelForCausalLM.from_pretrained(
"kevinkakkad168/earnings-alpha-llm",
torch_dtype=torch.float16,
device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("kevinkakkad168/earnings-alpha-llm")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
See the project GitHub for the full pipeline.
References
- Mayew & Venkatachalam (2012): The Power of Voice
- Li (2008): Annual report readability and earnings persistence
Quellen
1 Quelle
Quellen
1 QuelleVerifiziert 31. Okt.
Download files
Model · 4 weight files
All files listed in the provider index are present.
- model.safetensors.index.json · provider index
- config.json
- generation_config.json
- model-00004-of-00004.safetensors · individual shard
- model-00001-of-00004.safetensors · individual shard
- model-00002-of-00004.safetensors · individual shard
- model-00003-of-00004.safetensors · individual shard
Repository files
File inventory only; completeness is unverified.
Modellartefakte
4 Artefakte
Modellartefakte
4 Artefaktemodel-00001-of-00004.safetensors
safetensors · 4,63 GB · SHA-256 8db695a3d2f8…ff4d · Hugging Face
Download shardmodel-00002-of-00004.safetensors
safetensors · 4,66 GB · SHA-256 74bd42a19bab…b1c7 · Hugging Face
Download shardmodel-00003-of-00004.safetensors
safetensors · 4,58 GB · SHA-256 3c168a2db014…712e · Hugging Face
Download shardmodel-00004-of-00004.safetensors
safetensors · 1,09 GB · SHA-256 f4adabc4798b…62be · Hugging Face
Download shardQuellenauszüge
3 Auszüge
Quellenauszüge
3 Auszüge--- language: - en license: llama3.1 base_model: meta-llama/Meta-Llama-3.1-8B-Instruct tags: - finance - earnings - alpha - fine-tuned - lora --- # Earnings Call Alpha Extractor Fine-tuned Llama 3.1 8B for extracting quantitative signals from earnings call Q&A transcripts. ## What it does Scores earnings call Q&A sections on two dimensions: - **Confidence score** (1-10): How confident and direct is management? - **Evasiveness score** (1-10): How evasive, hedging, or deflecting is management? The mismatch between these scores and reported financial results (EPS surprise) forms an alpha signal with statistically significant IC. ## Training - **Base model**: `meta-llama/Meta-Llama-3.1-8B-Instruct` - **Fine-tuning**: LoRA on Together AI - **Training data**: ~700 earnings calls labeled by Llama 3.3 70B (Groq) - **Dataset**: Rogersurf/earnings-call-transcripts (9k+ calls, 2023-2026) ## Backtest results | Metric | Value | |--------|-------| | IC (5-day) | +0.076 | | IC t-stat | +2.00 (significant) | | Sharpe | 0.74 | | N events | 700 | | Period | 2023–2026 | ## Usage ```python from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline import torch model = AutoModelForCausalLM.from_pretrained( "kevinkakkad168/earnings-alpha-llm", torch_dtype=torch.float16, device_map="auto", ) tokenizer = AutoTokenizer.from_pretrained("kevinkakkad168/earnings-alpha-llm") pipe = pipeline("text-generation", model=model, tokenizer=tokenizer) ``` See the [project GitHub](https://github.com/your-username/earnings-alpha-extractor) for the full pipeline. ## References - Mayew & Venkatachalam (2012): *The Power of Voice* - Li (2008): *Annual report readability and earnings persistence*
Source context: 10 downloads · 0 likes · Repo kevinkakkad168/earnings-alpha-llm
Source context: 164 downloads · 0 likes · Repo kevinkakkad168/earnings-alpha-llm
