INDEX

Explanations

words related to the term "efficient" or similar concepts

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

 guiActiveUnfocused

-0.70

 Wolves

-0.68

 oath

-0.68

SHIP

-0.67

 Split

-0.66

 smartphones

-0.65

²¾

-0.64

meal

-0.63

 Chosen

-0.62

 Borderlands

-0.62

POSITIVE LOGITS

luent

1.41

emin

1.22

endi

1.21

iency

1.19

ence

1.16

erence

1.14

usion

1.14

usive

1.14

encing

1.13

erves

1.13

Activations Density 0.019%

words related to the term "efficient" or similar concepts

No Comments

No Known Activations