INDEX

Explanations

tips and recommendations regarding various aspects of writing and development, particularly related to technical concepts

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

alysed

-0.69

resy

-0.68

MH

-0.68

Bind

-0.67

ses

-0.64

ements

-0.64

aled

-0.63

iously

-0.62

olves

-0.62

ige

-0.61

POSITIVE LOGITS

 suppose

1.11

 imagine

0.95

 whereas

0.82

 consider

0.80

if

0.80

 Suppose

0.79

 although

0.75

avorite

0.71

 beware

0.69

 while

0.69

Activations Density 0.084%

tips and recommendations regarding various aspects of writing and development, particularly related to technical concepts

No Comments

No Known Activations