INDEX

Explanations

references to hacking or hacker-related terms

oai_token-act-pair · gpt-3.5-turbo

references to hacking and cybersecurity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Lago

-0.72

 citiz

-0.71

ufact

-0.69

 heirs

-0.65

uana

-0.65

etus

-0.64

hani

-0.63

Ã±a

-0.63

ttle

-0.62

 Seym

-0.61

POSITIVE LOGITS

tips

0.68

sels

0.68

 Dull

0.64

 Dirty

0.62

estern

0.62

fil

0.60

itiveness

0.60

cs

0.58

 mouse

0.57

nuts

0.57

Activations Density 1.011%

references to hacking or hacker-related terms

references to hacking and cybersecurity

No Comments

No Known Activations