INDEX

Explanations

references to a specific location or entity spelled as 'pool'

oai_token-act-pair · gpt-3.5-turbo

references to pools or groupings related to various contexts

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

rians

-0.83

rian

-0.74

Cel

-0.68

DonaldTrump

-0.68

eme

-0.67

NESS

-0.66

 Realms

-0.61

WTO

-0.60

 Loving

-0.60

======

-0.59

POSITIVE LOGITS

 pool

1.13

esville

1.10

pool

1.06

 Pool

1.00

 pools

0.95

regate

0.90

eries

0.88

side

0.87

erves

0.83

hare

0.79

Activations Density 0.023%

references to a specific location or entity spelled as 'pool'

references to pools or groupings related to various contexts

No Comments

No Known Activations