INDEX

Explanations

simple and straightforward statements or ideas

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

vance

-0.84

ashington

-0.81

 largeDownload

-0.79

hovah

-0.76

inburgh

-0.74

 extensively

-0.69

raints

-0.69

haw

-0.69

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.68

reon

-0.68

POSITIVE LOGITS

tons

1.34

minded

0.95

ton

0.92

 arithmetic

0.92

 minded

0.91

 syrup

0.88

wallet

0.87

 explanation

0.81

 elegance

0.79

 geometric

0.79

Activations Density 0.075%

simple and straightforward statements or ideas

No Comments

No Known Activations