INDEX

Explanations

adjective-noun pairs that express negative judgment or disapproval

oai_token-act-pair · gpt-3.5-turbo

terms related to disgrace and dishonor

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 ellipt

-0.75

 envelope

-0.71

 pores

-0.71

 composing

-0.70

oult

-0.70

 anecd

-0.69

 rink

-0.69

arten

-0.69

 parcels

-0.68

 zoning

-0.67

POSITIVE LOGITS

ful

1.27

fully

1.19

 disgrace

0.92

faced

0.92

ably

0.86

ously

0.78

forth

0.78

orable

0.78

à¤¾

0.77

ious

0.77

Activations Density 0.023%

adjective-noun pairs that express negative judgment or disapproval

terms related to disgrace and dishonor

No Comments

No Known Activations