INDEX

Explanations

the word "don" repeated in various contexts

oai_token-act-pair · gpt-3.5-turbo

the word "don" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 entitled

-0.62

 handic

-0.59

 airborne

-0.59

 observable

-0.59

 scheduled

-0.58

 grid

-0.57

 remotely

-0.57

 horn

-0.57

 marks

-0.55

 drastic

-0.55

POSITIVE LOGITS

don

4.57

DON

2.14

den

1.66

Don

1.58

DON

1.41

tons

1.36

do

1.30

didn

1.27

dos

1.24

die

1.21

Activations Density 0.010%

the word "don" repeated in various contexts

the word "don" in various contexts

No Comments

No Known Activations