INDEX

Explanations

phrases related to formal communication or sharing information

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

 reper

-0.78

Dyn

-0.61

fug

-0.61

 Kand

-0.60

Jah

-0.59

 Atomic

-0.58

 absor

-0.58

 pend

-0.57

Ow

-0.56

 Slater

-0.56

POSITIVE LOGITS

Disclaimer

0.83

advertising

0.78

Companies

0.77

Gay

0.75

Tweet

0.75

Share

0.75

 kindred

0.74

Hi

0.74

Bernie

0.74

Pinterest

0.73

Activations Density 0.047%

phrases related to formal communication or sharing information

No Comments

No Known Activations