INDEX

Explanations

mentions of events or products being released or announced

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

¬¼

-0.67

utive

-0.65

ļéĨĴ

-0.61

USS

-0.61

irie

-0.60

ãĤ´ãĥ³

-0.60

UF

-0.59

otin

-0.59

è£ħ

-0.58

herent

-0.58

POSITIVE LOGITS

 albeit

1.40

 namely

1.31

huh

1.29

but

1.27

 although

1.18

 though

1.17

eh

1.15

 especially

1.11

 whereas

1.11

 however

1.10

Activations Density 2.871%

mentions of events or products being released or announced

No Comments

No Known Activations