INDEX

Explanations

references to the term 'pit' in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to pit bull dogs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Lauder

-0.65

Pacific

-0.59

 Pacific

-0.55

kj

-0.54

 Marketplace

-0.54

 Empires

-0.53

 Spac

-0.53

 Feinstein

-0.52

eds

-0.52

POSITIVE LOGITS

iful

1.58

ifully

1.48

iless

1.48

cair

1.42

bull

1.24

iable

1.20

cher

1.06

uit

1.04

fall

1.01

falls

1.00

Activations Density 0.041%

references to the term 'pit' in various contexts

references to pit bull dogs

No Comments

No Known Activations