INDEX

Explanations

words related to law enforcement or legal processes

oai_token-act-pair · gpt-3.5-turbo

specific punctuation marks that appear frequently in the text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

geries

-0.79

 carbohyd

-0.77

 pyramid

-0.71

 satell

-0.70

 wagon

-0.69

ijah

-0.69

 decomp

-0.68

 fortun

-0.68

wana

-0.67

 accommodations

-0.67

POSITIVE LOGITS

ï¸ı

1.53

ï¸

0.97

 Thankfully

0.95

Thankfully

0.93

 Therefore

0.91

ðŁĺ

0.89

 Fortunately

0.88

Therefore

0.88

è£

0.85

âĶĢâĶĢâĶĢâĶĢ

0.85

Activations Density 0.143%

words related to law enforcement or legal processes

specific punctuation marks that appear frequently in the text

No Comments

No Known Activations