INDEX

Explanations

phrases related to causes or reasons

oai_token-act-pair · gpt-3.5-turbo

references to government and its influence

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

heit

-0.78

ateg

-0.77

achelor

-0.76

vier

-0.74

 describes

-0.71

adj

-0.71

itals

-0.71

soDeliveryDate

-0.71

aurus

-0.71

FIELD

-0.70

POSITIVE LOGITS

 absence

1.23

 presence

1.20

 inability

1.12

 fact

1.12

 generosity

1.10

 lack

1.04

 sheer

1.02

 influx

1.02

 arrival

1.00

 whims

0.99

Activations Density 0.280%

phrases related to causes or reasons

references to government and its influence

No Comments

No Known Activations