INDEX

Explanations

words related to defense and protection

oai_token-act-pair · gpt-3.5-turbo

instances of the term "defending" in various contexts

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

Scale

-0.85

STEM

-0.76

oscope

-0.76

isite

-0.73

 scale

-0.72

ement

-0.70

Exit

-0.69

Zip

-0.68

ue

-0.67

MRI

-0.66

POSITIVE LOGITS

 defending

3.84

 defend

2.07

 defended

1.90

 defends

1.81

 attacking

1.79

 protecting

1.68

 defense

1.58

 prosecuting

1.57

 guarding

1.56

 opposing

1.50

Activations Density 0.009%

words related to defense and protection

instances of the term "defending" in various contexts

No Comments

No Known Activations