INDEX

Explanations

words related to scientific research and medical terms

oai_token-act-pair · gpt-3.5-turbo

terms related to neuroscience and brain function

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Yep

-0.75

 anyway

-0.68

 Forget

-0.66

 buzzing

-0.64

Want

-0.64

Things

-0.63

SEE

-0.63

 Things

-0.62

 awfully

-0.62

 crap

-0.62

POSITIVE LOGITS

 differed

1.36

 yielded

1.30

 consisted

1.30

 exhibited

1.25

 demonstrated

1.21

 resulted

1.20

 showed

1.19

 occurred

1.07

 underwent

1.07

 appeared

1.05

Activations Density 0.320%

words related to scientific research and medical terms

terms related to neuroscience and brain function

No Comments

No Known Activations