INDEX

Explanations

details or additions being included in a text

oai_token-act-pair · gpt-3.5-turbo

phrases or expressions that indicate adding or including something

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

Consumer

-0.75

CHAT

-0.70

CBC

-0.70

Produ

-0.64

Business

-0.63

Application

-0.63

arters

-0.62

BSD

-0.62

termination

-0.61

icken

-0.61

POSITIVE LOGITS

 flair

0.96

 layer

0.94

 layers

0.87

 overhead

0.73

 thereto

0.72

 reinforcements

0.72

 suffix

0.71

 onto

0.70

 emphasis

0.70

 icing

0.70

Activations Density 0.147%

details or additions being included in a text

phrases or expressions that indicate adding or including something

No Comments

No Known Activations