INDEX

Explanations

user profiles on an online platform based on the number of posts they have made

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Posts" and associated numerical values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Bhar

-0.73

 Palest

-0.69

ONSORED

-0.66

fitted

-0.65

bid

-0.65

inant

-0.63

ãĥ¼ãĤ¯

-0.62

 Horowitz

-0.62

fitting

-0.61

itably

-0.61

POSITIVE LOGITS

 Posts

1.46

Posts

1.12

ertodd

0.96

 posts

0.90

 Joined

0.90

reply

0.82

 Posted

0.81

Joined

0.77

 Offense

0.77

################

0.74

Activations Density 0.008%

user profiles on an online platform based on the number of posts they have made

instances of the word "Posts" and associated numerical values

No Comments

No Known Activations