INDEX

Explanations

occurrences where a statement or action is being highlighted or emphasized

oai_token-act-pair · gpt-3.5-turbo

introductory phrases or transitions that indicate a statement or a point being made

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

EngineDebug

-0.63

udic

-0.58

],"

-0.57

',"

-0.56

Semitic

-0.53

igenous

-0.51

"}],"

-0.51

0000000000000000

-0.51

ãĥĺãĥ©

-0.50

querade

-0.49

POSITIVE LOGITS

cknowled

0.64

terness

0.56

cing

0.53

 Info

0.44

 Serving

0.44

 Recover

0.43

 kernels

0.42

 Symptoms

0.41

alyses

0.41

 Figures

0.41

Activations Density 0.766%

occurrences where a statement or action is being highlighted or emphasized

introductory phrases or transitions that indicate a statement or a point being made

No Comments

No Known Activations