INDEX

Explanations

symbols or patterns represented through a sequence of characters

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols 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 @ 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

 Samar

-0.67

itably

-0.67

orbit

-0.66

 scatter

-0.66

 scramble

-0.65

 Bengal

-0.62

 Brunswick

-0.61

 Bermuda

-0.61

 Triangle

-0.61

 cyan

-0.60

POSITIVE LOGITS

âĢ

0.90

¹

0.83

į

0.80

Â§

0.80

âĢł

0.78

Dialogue

0.78

acca

0.73

£

0.73

Ī

0.73

say

0.73

Activations Density 0.754%

symbols or patterns represented through a sequence of characters

special characters or symbols that appear frequently in the text

No Comments

No Known Activations