INDEX

Explanations

words related to analyzing or measuring technology or tools

oai_token-act-pair · gpt-3.5-turbo

words related to "analyzing" or "tests."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

unct

-0.61

 conventions

-0.61

 cape

-0.60

Vid

-0.59

ac

-0.59

 conformity

-0.58

 worship

-0.58

 spending

-0.58

 Passion

-0.58

Sac

-0.56

POSITIVE LOGITS

zer

4.74

zers

3.45

zing

2.02

zes

1.90

zed

1.88

ze

1.85

zen

1.68

zb

1.62

zy

1.60

zman

1.53

Activations Density 0.012%

words related to analyzing or measuring technology or tools

words related to "analyzing" or "tests."

No Comments

No Known Activations