INDEX

Explanations

the word "fore" followed by a high number

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ensable

-0.63

 Archdemon

-0.60

 Mine

-0.59

 disl

-0.59

 Nieto

-0.59

 toilets

-0.58

 Mechanics

-0.58

 Chic

-0.57

 bully

-0.57

 random

-0.57

POSITIVE LOGITS

shadow

1.49

told

1.43

sight

1.41

closed

1.37

gone

1.33

warn

1.33

nsics

1.32

runner

1.30

seen

1.28

warning

1.27

Activations Density 0.015%

the word "fore" followed by a high number

No Comments

No Known Activations