INDEX

Explanations

phrases related to interactions with specific individuals

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "whom."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Loading

-0.79

aster

-0.69

belt

-0.67

reach

-0.64

Charg

-0.63

 Delicious

-0.63

âĨ

-0.62

atory

-0.62

hands

-0.62

POSITIVE LOGITS

soever

2.01

omever

0.86

igham

0.80

ertodd

0.79

 distingu

0.79

ispers

0.78

 redes

0.78

etz

0.76

odox

0.76

vou

0.76

Activations Density 0.009%

phrases related to interactions with specific individuals

occurrences of the word "whom."

No Comments

No Known Activations