INDEX

Explanations

mentions or discussions of commonly known or referred information

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Integrity

-0.76

agate

-0.66

acity

-0.66

letico

-0.64

anth

-0.64

utenberg

-0.64

jit

-0.63

 Wish

-0.63

orate

-0.63

iants

-0.62

POSITIVE LOGITS

 referred

1.28

entimes

1.13

 regarded

1.11

 seen

1.10

 viewed

1.03

Used

1.02

 termed

1.01

 cited

1.01

 known

1.01

 associated

1.01

Activations Density 0.060%

mentions or discussions of commonly known or referred information

No Comments

No Known Activations