INDEX

Explanations

the word "non"

oai_token-act-pair · gpt-3.5-turbo

terms related to non-traditional or non-compliance concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Tycoon

-1.14

 Cobra

-0.84

è¦ļéĨĴ

-0.77

Recomm

-0.73

 Gardens

-0.73

 Armory

-0.70

 veins

-0.67

 nucleus

-0.66

 Mansion

-0.66

 Jaguars

-0.65

POSITIVE LOGITS

etheless

1.56

theless

1.03

profit

1.03

profits

1.03

withstanding

0.99

epad

0.97

partisan

0.96

onymous

0.94

jan

0.90

fiction

0.89

Activations Density 0.006%

the word "non"

terms related to non-traditional or non-compliance concepts

No Comments

No Known Activations