INDEX

Explanations

words containing the substring "reg"

oai_token-act-pair · gpt-3.5-turbo

references to regulations or regulatory concepts

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

 overhead

-0.64

 bondage

-0.64

Owl

-0.64

 Meow

-0.64

 firsthand

-0.63

 Wicked

-0.63

 ambush

-0.62

 Onion

-0.61

 peril

-0.61

 Danger

-0.61

POSITIVE LOGITS

reg

4.28

REG

2.10

Reg

1.93

regation

1.75

REG

1.65

Reg

1.46

regulated

1.37

region

1.35

register

1.34

regulation

1.33

Activations Density 0.009%

words containing the substring "reg"

references to regulations or regulatory concepts

No Comments

No Known Activations