INDEX

Explanations

phrases related to explaining or describing something in detail

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

ktop

-0.75

twitch

-0.73

tes

-0.73

uld

-0.70

imet

-0.69

 wisely

-0.66

talk

-0.63

Î±

-0.63

usb

-0.62

eka

-0.61

POSITIVE LOGITS

 similarities

0.69

 conformity

0.68

oran

0.65

 exemplary

0.65

tons

0.65

isance

0.62

 Heads

0.60

 cooperative

0.58

 executions

0.58

ãĥ£

0.57

Activations Density 0.163%

phrases related to explaining or describing something in detail

No Comments

No Known Activations