INDEX

Explanations

words related to specific knowledge or information

oai_token-act-pair · gpt-3.5-turbo

the word "Know" in various contexts

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

 suspension

-0.67

 turb

-0.58

wrapper

-0.58

 relegated

-0.58

 refreshing

-0.57

 gren

-0.57

res

-0.57

 normal

-0.57

 comes

-0.57

 cooled

-0.56

POSITIVE LOGITS

 Know

3.85

Know

2.46

 KNOW

2.04

know

1.91

 know

1.49

 Understand

1.36

 Knowledge

1.35

 Knowing

1.31

 Heard

1.24

 Hear

1.23

Activations Density 0.013%

words related to specific knowledge or information

the word "Know" in various contexts

No Comments

No Known Activations