INDEX

Explanations

mentions of different types of offices in various locations

oai_token-act-pair · gpt-3.5-turbo

mentions of "offices."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Adin

-0.77

Beg

-0.76

âķĲâķĲ

-0.72

oise

-0.71

rous

-0.70

ebook

-0.69

thia

-0.69

COMPLE

-0.67

oÄŁ

-0.67

Pause

-0.67

POSITIVE LOGITS

holder

0.90

holders

0.89

 offices

0.88

tops

0.87

office

0.82

 desks

0.78

matically

0.77

person

0.77

matical

0.77

 drawer

0.76

Activations Density 0.020%

mentions of different types of offices in various locations

mentions of "offices."

No Comments

No Known Activations