INDEX

Explanations

proper nouns related to different topics, such as works of fiction, politics, and technology

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

FTWARE

-0.66

Pac

-0.66

 benefited

-0.65

aceutical

-0.65

HTTP

-0.64

fill

-0.60

Owner

-0.58

alpha

-0.58

gans

-0.57

View

-0.57

POSITIVE LOGITS

 least

1.45

onement

1.16

yp

1.01

 halftime

1.00

mosp

0.98

abase

0.96

las

0.96

rium

0.95

roph

0.94

 dusk

0.92

Activations Density 0.935%

proper nouns related to different topics, such as works of fiction, politics, and technology

No Comments

No Known Activations