INDEX

Explanations

phrases related to achieving a goal or completing a task

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Tanz

-0.68

 Lies

-0.63

 cardinal

-0.63

 Inher

-0.62

iasco

-0.59

 defe

-0.58

iege

-0.58

alias

-0.57

 folly

-0.56

ynski

-0.56

POSITIVE LOGITS

rid

1.55

 acquainted

1.05

 away

0.94

 Started

0.94

 ahead

0.87

 reimb

0.87

 into

0.86

 INTO

0.86

 traction

0.85

 distracted

0.85

Activations Density 0.103%

phrases related to achieving a goal or completing a task

No Comments

No Known Activations