INDEX

Explanations

instances where "to" is followed by a verb or an adjective

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 reserves

-0.61

job

-0.59

 Known

-0.56

rieved

-0.55

 copied

-0.54

 samples

-0.54

 compressed

-0.54

 weighing

-0.53

 accommodations

-0.53

pict

-0.53

POSITIVE LOGITS

ggles

1.35

ilet

1.24

wered

1.19

pless

1.12

othy

1.07

asted

1.06

ads

1.06

asting

1.05

adies

1.02

asters

1.00

Activations Density 0.118%

instances where "to" is followed by a verb or an adjective

No Comments

No Known Activations