INDEX

Explanations

mentions of the term "rat"

oai_token-act-pair · gpt-3.5-turbo

references to rats

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

Kyl

-0.90

Kau

-0.76

 Odin

-0.72

duration

-0.72

 Legends

-0.70

vity

-0.69

stals

-0.68

 Flowers

-0.68

heim

-0.68

properties

-0.67

POSITIVE LOGITS

rat

3.87

Rat

2.15

 rats

1.67

 rodent

1.64

Rat

1.61

rat

1.47

 Rats

1.42

 ratification

1.34

 rabbit

1.28

 mouse

1.24

Activations Density 0.007%

mentions of the term "rat"

references to rats

No Comments

No Known Activations