INDEX

Explanations

names of universities and academic titles

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

ãĤ©

-0.87

alyst

-0.83

ocre

-0.75

urgical

-0.73

ceptive

-0.72

atform

-0.71

resso

-0.70

DEP

-0.70

urgy

-0.70

tools

-0.70

POSITIVE LOGITS

XVI

1.08

XIV

1.00

 Duchess

0.98

 throne

0.93

 Crown

0.89

 Queen

0.85

 Edinburgh

0.85

 Seym

0.83

 Anniversary

0.82

 Lauder

0.82

Activations Density 0.145%

names of universities and academic titles

No Comments

No Known Activations