INDEX

Explanations

expressions or exclamations of surprise or amazement

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

•layer 8 - surprise

No Comments

Negative Logits

Ö¼

-0.61

iership

-0.60

 manned

-0.59

ighthouse

-0.59

acco

-0.58

ourse

-0.58

Lod

-0.58

actionDate

-0.58

iffe

-0.58

orthy

-0.57

POSITIVE LOGITS

hhh

1.00

 yeah

0.99

ooo

0.93

hhhh

0.93

oooo

0.92

ahah

0.91

yea

0.89

oh

0.88

 kidding

0.86

oooooooo

0.85

Activations Density 0.121%

expressions or exclamations of surprise or amazement

No Comments

No Known Activations