INDEX

Explanations

numerical values or measurements in a text

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 condem

-0.65

 disposed

-0.65

 redesign

-0.64

dayName

-0.63

 dstg

-0.62

 agre

-0.60

 performance

-0.56

 lapt

-0.56

ertodd

-0.54

leep

-0.53

POSITIVE LOGITS

 those

0.83

us

0.80

 these

0.79

them

0.79

those

0.79

course

0.76

 them

0.76

 course

0.73

our

0.72

these

0.71

Activations Density 2.841%

numerical values or measurements in a text

No Comments

No Known Activations