INDEX

Explanations

phrases related to philosophical or legal discussions

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

kay

-0.77

alls

-0.73

 Kush

-0.72

 yacht

-0.71

bush

-0.71

 Kass

-0.71

kees

-0.71

 Kane

-0.71

 sausage

-0.70

kj

-0.69

POSITIVE LOGITS

vernment

0.84

EngineDebug

0.81

 supra

0.75

ational

0.75

essional

0.72

piracy

0.72

reme

0.71

ULT

0.71

oreal

0.69

ternal

0.69

Activations Density 1.187%

phrases related to philosophical or legal discussions

No Comments

No Known Activations