INDEX

Explanations

brand names and company names

oai_token-act-pair · gpt-3.5-turbo

the ampersand symbol '&'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

etheless

-1.08

 misunder

-0.79

atche

-0.75

Downloadha

-0.73

 Leilan

-0.73

ionage

-0.72

 Rohing

-0.72

nces

-0.70

reddits

-0.70

 constitu

-0.70

POSITIVE LOGITS

#$

0.91

vine

0.86

NAS

0.84

++++++++++++++++

0.81

 Associates

0.79

 vice

0.78

GAN

0.77

etc

0.76

then

0.74

optional

0.72

Activations Density 0.064%

brand names and company names

the ampersand symbol '&'

No Comments

No Known Activations