INDEX

Explanations

references to different types of hairstyles

oai_token-act-pair · gpt-3.5-turbo

terms related to hairstyles and infringement

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

 Trip

-0.70

 Olympic

-0.69

 captive

-0.67

 River

-0.66

 Coat

-0.65

 Metropolitan

-0.63

 Roman

-0.62

ibaba

-0.62

 Hudson

-0.61

 Camden

-0.61

POSITIVE LOGITS

Īè

1.02

ĭ

0.96

accompan

0.94

ritten

0.92

elines

0.92

yth

0.90

¶æ

0.89

wip

0.88

shown

0.87

£

0.86

Activations Density 0.028%

references to different types of hairstyles

terms related to hairstyles and infringement

No Comments

No Known Activations