INDEX

Explanations

phrases related to political statements or discussions

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

 shroud

-0.79

 couch

-0.68

 semblance

-0.67

 Belg

-0.65

 hemor

-0.63

 Doodle

-0.63

 taxp

-0.62

 guiActiveUnfocused

-0.62

entary

-0.62

 canvas

-0.62

POSITIVE LOGITS

ª

1.28

¹

1.23

¸

1.10

ł

1.08

Ĳ

1.08

ĳ

1.06

ı

1.03

¤

1.01

¡

1.00

³

0.99

Activations Density 0.154%

phrases related to political statements or discussions

No Comments

No Known Activations