INDEX

Explanations

specific names related to literature, cinema, and media

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

IST

-0.77

 Inquis

-0.69

 Equip

-0.69

ocamp

-0.66

 Mutual

-0.65

opathy

-0.64

enegger

-0.63

sburgh

-0.63

ĸļ

-0.62

 guiActiveUnfocused

-0.62

POSITIVE LOGITS

bles

1.43

bling

1.37

bled

1.34

bler

1.13

plet

1.12

blers

1.10

bt

1.07

rier

1.06

plets

1.02

pees

0.98

Activations Density 0.019%

specific names related to literature, cinema, and media

No Comments

No Known Activations