INDEX

Explanations

phrases related to intense and often negative emotional responses

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols used 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 scatter

-0.66

 cyan

-0.56

dig

-0.56

wagen

-0.56

anwhile

-0.54

 scattering

-0.53

 detached

-0.53

 radar

-0.52

 relocation

-0.51

 Bermuda

-0.51

POSITIVE LOGITS

¹

0.84

¬

0.79

£

0.76

į

0.75

âĹ¼

0.74

ı

0.74

catentry

0.73

¼

0.72

º

0.72

Â¢

0.68

Activations Density 0.586%

phrases related to intense and often negative emotional responses

special characters or symbols used in the text

No Comments

No Known Activations