INDEX

Explanations

positive emotional reactions

oai_token-act-pair · gpt-3.5-turbo

expressions of surprise or awe

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Townsend

-0.64

 leaflets

-0.64

 Feld

-0.63

 Fract

-0.60

Vie

-0.59

 Nadu

-0.59

 discriminated

-0.59

 duplication

-0.58

 inadvertently

-0.58

 Missile

-0.58

POSITIVE LOGITS

kward

1.57

akening

1.45

akens

1.38

aken

1.18

esome

1.18

FUL

1.06

oken

1.01

ait

1.00

oln

0.99

ful

0.99

Activations Density 0.020%

positive emotional reactions

expressions of surprise or awe

No Comments

No Known Activations