INDEX

Explanations

information related to publications, authors, and studies

oai_token-act-pair · gpt-3.5-turbo

references to published works and their details

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 bush

-0.63

Distance

-0.58

ahime

-0.57

rugged

-0.53

 choking

-0.53

?'

-0.51

«

-0.50

hawks

-0.50

existent

-0.49

 flung

-0.49

POSITIVE LOGITS

 archive

0.72

xtap

0.68

 excerpts

0.68

 anonym

0.64

 archives

0.64

 publishing

0.63

 annotations

0.63

 publication

0.63

 redacted

0.62

 acron

0.62

Activations Density 2.067%

information related to publications, authors, and studies

references to published works and their details

No Comments

No Known Activations