INDEX

Explanations

quotes or dialogues from people discussing various topics

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

antine

-0.68

 joints

-0.67

utsche

-0.60

 Holo

-0.60

izon

-0.60

sterdam

-0.59

oteric

-0.58

 patents

-0.56

 subsidized

-0.56

mbuds

-0.56

POSITIVE LOGITS

Narr

0.91

Said

0.83

'"

0.83

âĹ¼

0.81

Asked

0.80

]"

0.80

Refer

0.79

Shape

0.78

Ö¼

0.77

>>>

0.77

Activations Density 0.573%

quotes or dialogues from people discussing various topics

No Comments

No Known Activations