INDEX

Explanations

health-related information and statistics

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

orem

-0.80

amus

-0.68

enary

-0.67

rup

-0.66

SpaceEngineers

-0.65

eur

-0.65

inea

-0.65

oba

-0.64

WARN

-0.64

atorium

-0.64

POSITIVE LOGITS

 including

1.86

 namely

1.67

 includ

1.49

 Including

1.47

 ranging

1.46

including

1.44

 notably

1.39

 culminating

1.18

 albeit

1.16

 totaling

1.12

Activations Density 0.448%

health-related information and statistics

No Comments

No Known Activations