INDEX

Explanations

phrases that conclude a text or an argument

oai_token-act-pair · gpt-3.5-turbo

a specific repeated word or term, possibly indicating emphasis or significance in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ription

-0.67

 outline

-0.66

 unim

-0.64

 maximum

-0.63

 ultras

-0.61

ufact

-0.60

 stri

-0.60

 manifesto

-0.59

 settlements

-0.59

 integ

-0.59

POSITIVE LOGITS

Ŀ

1.08

ļ

0.94

º

0.87

bryce

0.84

£

0.84

çľ

0.83

Ĭ

0.82

ĩ

0.81

¼

0.81

»

0.80

Activations Density 0.129%

phrases that conclude a text or an argument

a specific repeated word or term, possibly indicating emphasis or significance in the text

No Comments

No Known Activations