INDEX

Explanations

phrases related to legal cases or rulings

oai_token-act-pair · gpt-3.5-turbo

negative terms or concepts related to law, finance, or other formal processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

£ı

-0.57

ecause

-0.55

 tremend

-0.55

olulu

-0.53

 notor

-0.53

awaru

-0.53

raints

-0.52

iage

-0.51

resso

-0.50

Pwr

-0.50

POSITIVE LOGITS

2.16

âĢĳ

1.45

âĢĲ

1.36

-,

1.30

-$

1.21

-'

1.20

"-

1.13

-[

1.12

-.

1.08

'-

1.08

Activations Density 0.486%

phrases related to legal cases or rulings

negative terms or concepts related to law, finance, or other formal processes

No Comments

No Known Activations