INDEX

Explanations

rankings and scores

oai_token-act-pair · gpt-3.5-turbo

rankings and positions in lists or categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

umb

-0.92

gery

-0.76

................................................................

-0.70

 spilled

-0.68

 leased

-0.67

 delivered

-0.63

 attendant

-0.63

 vessels

-0.63

dayName

-0.62

 stretching

-0.62

POSITIVE LOGITS

 rankings

1.15

 ranking

1.13

ranked

1.09

 contender

1.06

 ranked

1.05

 honorable

1.03

 contenders

1.01

Top

0.99

rank

0.99

Top

0.97

Activations Density 0.930%

rankings and scores

rankings and positions in lists or categories

No Comments

No Known Activations