INDEX

Explanations

the word 'Description' or variations of it

oai_token-act-pair · gpt-3.5-turbo

references to descriptions of entities or concepts

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

ecake

-0.67

lift

-0.67

 Roosevelt

-0.65

 Wonderland

-0.63

 Fenrir

-0.63

 spare

-0.62

strap

-0.61

Reviewer

-0.61

buck

-0.61

Bir

-0.61

POSITIVE LOGITS

ript

1.37

desc

1.17

ribing

1.13

ribe

1.05

 Desc

1.05

ribed

1.03

ribes

1.02

endants

0.99

 desc

0.96

Desc

0.92

Activations Density 0.010%

the word 'Description' or variations of it

references to descriptions of entities or concepts

No Comments

No Known Activations