INDEX

Explanations

I'm sorry, but based on the activations provided, I'm unable to determine a specific pattern or theme that neuron 4 is looking for in the text

oai_token-act-pair · gpt-3.5-turbo

special characters or non-standard symbols 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

-0.91

raints

-0.72

 background

-0.70

 distracted

-0.69

 foreground

-0.69

 offending

-0.68

 plain

-0.68

gery

-0.67

 Plain

-0.67

 dividing

-0.67

POSITIVE LOGITS

Ħ

1.41

ĳ

1.14

¸

1.10

ļ

1.06

ĸ

1.06

Ð¸

1.03

Ĺ

1.03

¶

1.02

¼

1.01

¾

1.00

Activations Density 0.003%

I'm sorry, but based on the activations provided, I'm unable to determine a specific pattern or theme that neuron 4 is looking for in the text

special characters or non-standard symbols in the text

No Comments

No Known Activations