INDEX

Explanations

years

oai_token-act-pair · gpt-3.5-turbo

years mentioned in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

bors

-0.78

 notor

-0.76

lass

-0.76

nery

-0.75

imore

-0.74

bor

-0.74

onet

-0.73

aws

-0.72

abwe

-0.72

imer

-0.72

POSITIVE LOGITS

å¹

1.16

 edition

0.79

 INTER

0.76

 Edition

0.75

 EDITION

0.74

 SPACE

0.74

 Remastered

0.73

 Presidential

0.69

 Crimson

0.67

Activations Density 0.103%

years

years mentioned in the document

No Comments

No Known Activations