INDEX

Explanations

phrases prompting the reader to pay attention or consider something

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

lees

-0.87

oing

-0.80

sbm

-0.79

soever

-0.74

iar

-0.74

=~=~

-0.74

oided

-0.70

iere

-0.68

raged

-0.68

ittle

-0.66

POSITIVE LOGITS

WHY

1.11

 something

1.08

 what

1.05

why

1.04

how

0.99

 causation

0.97

 ourselves

0.96

 basics

0.94

 some

0.91

 facts

0.90

Activations Density 0.218%

phrases prompting the reader to pay attention or consider something

No Comments

No Known Activations