INDEX

Explanations

phrases where someone talks about what they like to do or what activities they enjoy

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

river

-0.63

rising

-0.63

ratulations

-0.63

abiding

-0.61

teen

-0.60

cheon

-0.60

activation

-0.60

isite

-0.60

rance

-0.58

 Settlement

-0.58

POSITIVE LOGITS

 remind

1.17

 speculate

1.14

 pretend

1.09

 think

1.09

 imagine

1.06

 visualize

1.04

 tease

1.04

 joke

1.03

 gamble

1.00

 hear

0.98

Activations Density 0.061%

phrases where someone talks about what they like to do or what activities they enjoy

No Comments

No Known Activations