INDEX

Explanations

topics related to anticipation or speculation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

elta

-0.76

endas

-0.72

©¶æ

-0.69

 Doodle

-0.69

 throats

-0.64

oop

-0.63

arus

-0.61

ses

-0.61

æĸ

-0.59

ingo

-0.58

POSITIVE LOGITS

ional

0.85

rance

0.77

rack

0.73

weight

0.70

rad

0.69

arian

0.65

risome

0.64

jee

0.63

 mortals

0.63

fact

0.62

Activations Density 0.022%

topics related to anticipation or speculation

No Comments

No Known Activations