INDEX

Explanations

the word "loved" or variations of it

oai_token-act-pair · gpt-3.5-turbo

words related to love and affection

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

 reorgan

-0.72

 lapse

-0.68

STATE

-0.67

 Blazers

-0.65

 flashes

-0.64

 rise

-0.62

Frameworks

-0.61

 scope

-0.61

 clear

-0.60

break

-0.59

POSITIVE LOGITS

oved

1.09

oving

0.94

ove

0.94

oves

0.93

nesday

0.87

iev

0.83

icz

0.81

ierrez

0.80

ative

0.79

tsky

0.79

Activations Density 0.011%

the word "loved" or variations of it

words related to love and affection

No Comments

No Known Activations