INDEX

Explanations

directives or instructions

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

±

-0.70

encer

-0.67

Downloadha

-0.66

¨

-0.65

¥µ

-0.65

anship

-0.64

abella

-0.64

istant

-0.63

akes

-0.62

ifled

-0.62

POSITIVE LOGITS

:-

1.41

:[

1.38

:"

1.36

1.30

:(

1.18

]:

1.12

():

1.08

:#

1.06

%:

1.05

:'

1.05

Activations Density 1.022%

directives or instructions

No Comments

No Known Activations