INDEX

Explanations

names written in short form

oai_token-act-pair · gpt-3.5-turbo

variations of the suffixes 'ht', 'ld', and 'gd' in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

zzo

-0.67

 slack

-0.62

fri

-0.60

ult

-0.54

 priceless

-0.54

================

-0.54

Fi

-0.54

 knockout

-0.53

âĢ¦âĢ¦âĢ¦âĢ¦

-0.53

pac

-0.53

POSITIVE LOGITS

opol

0.95

orf

0.93

ude

0.88

rums

0.87

ragon

0.85

eger

0.85

oche

0.84

rek

0.84

orp

0.83

irect

0.80

Activations Density 0.125%

names written in short form

variations of the suffixes 'ht', 'ld', and 'gd' in text

No Comments

No Known Activations