INDEX

Explanations

the word 'and'

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

>[

-0.82

eks

-0.77

Deal

-0.75

bugs

-0.73

cells

-0.65

obiles

-0.64

oused

-0.64

bytes

-0.63

sticks

-0.63

bage

-0.62

POSITIVE LOGITS

rogen

0.90

rogens

0.78

 thus

0.72

 therefore

0.71

 hence

0.70

0.69

 thereby

0.67

 Caribbean

0.67

ifice

0.66

 Caucasus

0.65

Activations Density 0.170%

the word 'and'

No Comments

No Known Activations