INDEX

Explanations

phrases related to taking action or making decisions

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

externalActionCode

-0.81

 Examiner

-0.65

yan

-0.63

Legendary

-0.60

uary

-0.60

 Associated

-0.58

taker

-0.56

}{

-0.56

 Correspond

-0.56

alias

-0.54

POSITIVE LOGITS

 aside

1.09

 forth

0.95

rescent

0.94

 together

0.90

rid

0.88

tin

0.83

 brakes

0.82

tering

0.81

atively

0.78

 pressure

0.77

Activations Density 3.190%

phrases related to taking action or making decisions

No Comments

No Known Activations