INDEX

Explanations

keywords related to providing information or insight

oai_token-act-pair · gpt-3.5-turbo

action verbs that indicate functionality or contribution within various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tone

-0.64

zo

-0.64

volent

-0.61

ozo

-0.61

nex

-0.60

 introduction

-0.59

street

-0.59

rone

-0.59

bearing

-0.56

Ø§ÙĦ

-0.56

POSITIVE LOGITS

paces

1.07

hift

1.06

ometimes

1.04

ettings

0.97

omething

0.93

hirt

0.90

creen

0.88

ilver

0.87

heet

0.84

pace

0.82

Activations Density 0.491%

keywords related to providing information or insight

action verbs that indicate functionality or contribution within various contexts

No Comments

No Known Activations