INDEX

Explanations

ambiguous or unclear words and phrases

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

 sidew

-0.72

 heel

-0.70

hement

-0.69

 intuitive

-0.67

 defamation

-0.67

 suppressed

-0.66

 sway

-0.66

 flush

-0.66

 disg

-0.65

 positively

-0.64

POSITIVE LOGITS

lihood

1.00

Else

0.93

âĦ¢

0.89

 Limits

0.86

tons

0.84

ness

0.83

itarian

0.82

!,

0.80

cott

0.79

sburg

0.79

Activations Density 0.162%

ambiguous or unclear words and phrases

No Comments

No Known Activations