INDEX

Explanations

URLs and code snippets

oai_token-act-pair · gpt-3.5-turbo

words related to medical and scientific terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 hawk

-0.70

redo

-0.68

 vain

-0.67

 tame

-0.65

 Krishna

-0.65

 Accountability

-0.65

Wiki

-0.64

 bondage

-0.64

 misdemeanor

-0.63

Sharp

-0.63

POSITIVE LOGITS

ocol

2.36

igl

2.01

opl

1.73

cill

1.44

ogl

1.36

ipl

1.36

vill

1.31

obl

1.20

 Panc

1.15

uve

1.14

Activations Density 0.041%

URLs and code snippets

words related to medical and scientific terminology

No Comments

No Known Activations