INDEX

Explanations

phrases related to products or services

oai_token-act-pair · gpt-3.5-turbo

names of products and items associated with packaging and testing

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

rising

-0.82

ischer

-0.76

lie

-0.71

fortunately

-0.71

 democr

-0.69

 acknow

-0.68

taboola

-0.68

azon

-0.68

etheless

-0.68

 guiActiveUn

-0.67

POSITIVE LOGITS

 Manager

1.20

 Replacement

1.14

 Calculator

1.11

 Agreement

1.10

 Collection

1.08

 Definition

1.07

 Holder

1.06

 Implementation

1.06

 Suite

1.05

 Facility

1.05

Activations Density 0.359%

phrases related to products or services

names of products and items associated with packaging and testing

No Comments

No Known Activations