INDEX

Explanations

terms related to technology, networks, and computer programming

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of institutions, organizations, or titles

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

 eleph

-1.01

ccording

-0.98

 ende

-0.89

 pione

-0.81

assetsadobe

-0.78

 exting

-0.77

 conclud

-0.77

Þ

-0.75

ñ

-0.74

 tremend

-0.73

POSITIVE LOGITS

verse

0.64

istar

0.63

inx

0.63

odo

0.63

nex

0.61

ceive

0.60

opt

0.60

olve

0.60

plet

0.60

isco

0.59

Activations Density 0.247%

terms related to technology, networks, and computer programming

proper nouns, particularly names of institutions, organizations, or titles

No Comments

No Known Activations