INDEX

Explanations

proper nouns or specific terms

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

srf

-0.59

 Drift

-0.54

 seas

-0.53

 drift

-0.53

tnc

-0.53

bledon

-0.52

lished

-0.52

 widest

-0.51

 tumble

-0.51

 descent

-0.50

POSITIVE LOGITS

conom

1.23

galitarian

0.92

theless

0.91

lect

0.90

ffect

0.85

essee

0.81

except

0.81

uther

0.80

lust

0.80

 whatsoever

0.80

Activations Density 0.015%

proper nouns or specific terms

No Comments

No Known Activations