INDEX

Explanations

dialogue interactions 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_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

icipated

-0.76

inement

-0.73

actionDate

-0.70

result

-0.68

inction

-0.68

 rall

-0.65

conservancy

-0.64

İĭ

-0.64

¥ŀ

-0.63

arcity

-0.63

POSITIVE LOGITS

 yeah

1.38

yeah

1.32

 Yeah

1.24

 kidding

1.20

sir

1.17

hhh

1.14

hhhh

1.12

Yeah

1.11

yea

1.08

 fuck

1.07

Activations Density 4.530%

dialogue interactions between characters

No Comments

No Known Activations