INDEX

Explanations

terms related to classification and categorization

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

 stump

-0.68

hiba

-0.68

aukee

-0.67

wit

-0.66

foreseen

-0.63

SIGN

-0.62

 spur

-0.62

OPLE

-0.60

edia

-0.59

²¾

-0.59

POSITIVE LOGITS

ifications

1.39

ifiers

1.30

ifier

1.29

ifies

1.21

ifying

1.20

ifiable

1.08

ification

1.02

ified

1.02

ically

1.01

mates

0.93

Activations Density 0.528%

terms related to classification and categorization

No Comments

No Known Activations