INDEX

Explanations

security-related terms or concepts

oai_token-act-pair · gpt-3.5-turbo

terms related to the concept of protection

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

ãĥ£

-0.79

bold

-0.67

ellar

-0.64

nexus

-0.63

umo

-0.63

RM

-0.63

LINE

-0.62

ffe

-0.61

zos

-0.60

ching

-0.60

POSITIVE LOGITS

ively

1.00

 afforded

0.91

iveness

0.84

 racket

0.82

folios

0.81

dogs

0.79

ously

0.78

raints

0.78

orship

0.76

 protection

0.75

Activations Density 0.029%

security-related terms or concepts

terms related to the concept of protection

No Comments

No Known Activations