INDEX

Explanations

phrases related to actions and activities over a period of time

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

apo

-0.73

 Tycoon

-0.65

lies

-0.65

MJ

-0.64

next

-0.64

poses

-0.61

rish

-0.61

 Later

-0.59

article

-0.59

Must

-0.59

POSITIVE LOGITS

 lately

1.31

 since

1.20

 steadily

1.02

 extensively

1.01

 repeatedly

0.94

 recently

0.94

 countless

0.90

since

0.88

 unsuccessfully

0.87

 hitherto

0.86

Activations Density 4.411%

phrases related to actions and activities over a period of time

No Comments

No Known Activations