INDEX

Explanations

phrases related to defense mechanisms and security measures

oai_token-act-pair · gpt-3.5-turbo

terms related to defenses and offenses

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Shape

-0.68

 Member

-0.68

 Nobel

-0.67

MLA

-0.65

 OECD

-0.61

NZ

-0.61

 Kazakh

-0.60

 travelling

-0.59

cil

-0.59

ient

-0.59

POSITIVE LOGITS

 defenses

3.89

 defences

3.19

 offenses

1.46

 weaknesses

1.07

 foes

1.04

 shields

1.03

 defenders

1.03

 protections

1.02

 defense

1.02

 circuits

1.01

Activations Density 0.014%

phrases related to defense mechanisms and security measures

terms related to defenses and offenses

No Comments

No Known Activations