INDEX

Explanations

dates and locations in news articles

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

quist

-0.69

ruct

-0.65

 Import

-0.64

Ess

-0.64

Secondly

-0.63

Planet

-0.61

Gender

-0.61

Rub

-0.60

âĹ¼

-0.60

Manufact

-0.59

POSITIVE LOGITS

phy

0.97

piring

0.94

pires

0.90

pired

0.88

 well

0.87

 soon

0.86

 evidenced

0.82

 opposed

0.81

 they

0.80

criptions

0.80

Activations Density 0.129%

dates and locations in news articles

No Comments

No Known Activations