INDEX

Explanations

complimentary phrases or words

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

ogens

-0.81

NF

-0.72

orig

-0.71

redit

-0.71

 Administ

-0.70

abad

-0.70

igate

-0.70

onics

-0.70

abies

-0.70

eligible

-0.69

POSITIVE LOGITS

 little

0.89

 breeze

0.84

 nice

0.82

 fluffy

0.80

 neat

0.79

bye

0.79

bye

0.79

 gesture

0.79

Ã±o

0.77

 sounding

0.77

Activations Density 0.041%

complimentary phrases or words

No Comments

No Known Activations