INDEX

Explanations

mentions of personalized content such as logins, emails, and tasks

oai_token-act-pair · gpt-3.5-turbo

the word "your" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

apo

-0.94

forth

-0.78

 Cohn

-0.75

 Goes

-0.72

Lago

-0.71

 Originally

-0.68

 Shapiro

-0.66

 Epstein

-0.66

wik

-0.65

aways

-0.64

POSITIVE LOGITS

own

1.40

 favourite

1.17

 favorite

1.07

 adversary

0.94

anmar

0.93

ocard

0.92

 desired

0.89

 opponent

0.89

 preferred

0.89

 imagination

0.88

Activations Density 0.105%

mentions of personalized content such as logins, emails, and tasks

the word "your" and its variations in different contexts

No Comments

No Known Activations