INDEX

Explanations

phrases related to a specific entity or name, likely "Sec9"

oai_token-act-pair · gpt-3.5-turbo

the term "Sec" followed by numerical values, indicating a focus on security-related references or designations

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

ortal

-0.72

 imitate

-0.70

mares

-0.69

alle

-0.67

amura

-0.64

atten

-0.64

ei

-0.63

atsuki

-0.62

landish

-0.62

axe

-0.62

POSITIVE LOGITS

Sec

4.02

Sec

2.57

sec

1.93

sec

1.75

SEC

1.33

SEC

1.28

 Sect

1.25

 Secure

1.05

 Sanct

0.96

 Priv

0.95

Activations Density 0.014%

phrases related to a specific entity or name, likely "Sec9"

the term "Sec" followed by numerical values, indicating a focus on security-related references or designations

No Comments

No Known Activations