INDEX

Explanations

concepts related to legal or institutional separations

oai_token-act-pair · gpt-3.5-turbo

references to the concept of separation, particularly in political or religious contexts

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

 behalf

-0.66

 Hels

-0.65

 Prospect

-0.65

verning

-0.65

TL

-0.64

si

-0.64

eon

-0.63

Lob

-0.62

enegger

-0.62

eb

-0.62

POSITIVE LOGITS

 separation

1.35

 separ

1.01

aration

0.81

separ

0.81

 separating

0.79

 separated

0.75

 headaches

0.75

 immersion

0.73

lement

0.73

 distances

0.73

Activations Density 0.005%

concepts related to legal or institutional separations

references to the concept of separation, particularly in political or religious contexts

No Comments

No Known Activations