INDEX

Explanations

words related to location or direction

oai_token-act-pair · gpt-3.5-turbo

variations of the word "somewhat."

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

ãĥ¼ãĥĨ

-0.80

DOC

-0.69

 Nielsen

-0.66

 Mongolia

-0.64

 Blaz

-0.64

DOC

-0.63

Reviewer

-0.63

STER

-0.61

 Metro

-0.61

 Gemini

-0.60

POSITIVE LOGITS

hat

1.64

here

1.34

hing

1.20

hest

1.04

hooting

1.03

heres

1.01

orld

0.99

actory

0.98

ravel

0.97

hin

0.97

Activations Density 0.049%

words related to location or direction

variations of the word "somewhat."

No Comments

No Known Activations