INDEX

Explanations

text sequences starting with "Because" followed by a reason or explanation

oai_token-act-pair · gpt-3.5-turbo

statements introducing explanations or justifications

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

ye

-0.73

thro

-0.70

resident

-0.70

mint

-0.69

Sham

-0.69

scr

-0.69

yan

-0.69

åĤ

-0.68

shr

-0.65

jer

-0.64

POSITIVE LOGITS

endment

0.87

urers

0.82

xual

0.79

akening

0.78

uristic

0.75

pite

0.73

*/(

0.72

rike

0.71

awaru

0.71

ecause

0.70

Activations Density 0.050%

text sequences starting with "Because" followed by a reason or explanation

statements introducing explanations or justifications

No Comments

No Known Activations