INDEX

Explanations

punctuations in combination with short phrases

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

irie

-0.80

enhagen

-0.74

gang

-0.73

robe

-0.72

enaries

-0.71

ument

-0.67

escription

-0.65

ioxide

-0.64

everal

-0.64

ļéĨĴ

-0.64

POSITIVE LOGITS

huh

1.23

 especially

1.10

 considering

1.07

 albeit

1.03

eh

1.02

especially

0.99

but

0.99

 though

0.97

 Especially

0.93

 although

0.92

Activations Density 0.225%

punctuations in combination with short phrases

No Comments

No Known Activations