INDEX

Explanations

phrases describing relationships between characters

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

strate

-0.87

ifies

-0.75

iband

-0.75

thood

-0.73

udo

-0.71

FIELD

-0.70

heit

-0.69

soever

-0.69

zac

-0.69

owned

-0.68

POSITIVE LOGITS

 absence

1.21

 fact

1.20

 sheer

1.19

 presence

1.15

 possibility

1.13

 latter

1.12

 extent

1.10

 aforementioned

1.10

 severity

1.06

 seriousness

1.05

Activations Density 0.307%

phrases describing relationships between characters

No Comments

No Known Activations