INDEX

Explanations

phrases related to reasons or causes

oai_token-act-pair · gpt-3.5-turbo

phrases reflecting reasons or justifications, particularly those related to the absence of reasons

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

eer

-0.69

heid

-0.66

Express

-0.65

dar

-0.63

ommel

-0.61

Files

-0.60

egg

-0.60

 drafts

-0.59

eers

-0.58

 Presence

-0.58

POSITIVE LOGITS

 reason

1.67

 reasons

1.65

 sake

1.51

 Reasons

1.18

 purposes

1.14

 Reason

1.01

 purpose

1.00

Reason

1.00

reason

0.95

 understandable

0.81

Activations Density 0.175%

phrases related to reasons or causes

phrases reflecting reasons or justifications, particularly those related to the absence of reasons

No Comments

No Known Activations