INDEX

Explanations

phrases related to personal opinions or reactions

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize personal opinions or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

neau

-0.72

ãĤ´ãĥ³

-0.70

ocaust

-0.70

imeter

-0.68

fee

-0.67

estones

-0.65

oided

-0.63

ammers

-0.62

enne

-0.62

orthy

-0.61

POSITIVE LOGITS

why

1.44

 what

1.28

how

1.24

 where

1.05

WHY

1.00

 whats

1.00

what

0.97

 exactly

0.91

why

0.89

 gonna

0.80

Activations Density 0.105%

phrases related to personal opinions or reactions

phrases that emphasize personal opinions or statements

No Comments

No Known Activations