INDEX

Explanations

the word "once" in various contexts

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

 nuances

-0.60

 weights

-0.57

 averaging

-0.56

 intern

-0.56

 incent

-0.54

 interpre

-0.53

 deviations

-0.52

 compress

-0.52

 exceptions

-0.51

 inserts

-0.50

POSITIVE LOGITS

again

1.02

 again

0.88

ESA

0.72

 Again

0.70

Fear

0.69

ueller

0.69

BUS

0.68

bare

0.66

moil

0.65

é¾įåĸļå£«

0.65

Activations Density 0.377%

the word "once" in various contexts

No Comments

No Known Activations