INDEX

Explanations

references to specific entities or terms having "rat" in them

oai_token-act-pair · gpt-3.5-turbo

instances of the word "rat" or variations of it

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

Negative Logits

 Fields

-0.69

 deploy

-0.66

 Leilan

-0.65

ggies

-0.60

KS

-0.60

 entry

-0.58

 deploying

-0.58

ãģį

-0.57

 Lily

-0.57

 Waterloo

-0.57

POSITIVE LOGITS

rat

4.64

rats

2.47

Rat

2.07

rates

1.84

rate

1.57

Rat

1.56

rating

1.52

rad

1.46

rations

1.37

ras

1.25

Activations Density 0.015%

references to specific entities or terms having "rat" in them

instances of the word "rat" or variations of it

No Known Activations