INDEX

Explanations

phrases indicating expectation or prediction

oai_token-act-pair · gpt-3.5-turbo

instances of the word "would" indicating speculation or future intentions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

holder

-0.68

meta

-0.67

è¦ļéĨĴ

-0.67

DAQ

-0.63

values

-0.63

æ³

-0.62

 Modified

-0.62

building

-0.61

isted

-0.58

DNA

-0.58

POSITIVE LOGITS

be

1.14

 inevitably

1.07

 never

0.98

 surely

0.96

 eventually

0.95

 suffice

0.94

 gladly

0.91

 undoubtedly

0.89

 invariably

0.88

 unleash

0.88

Activations Density 0.140%

phrases indicating expectation or prediction

instances of the word "would" indicating speculation or future intentions

No Comments

No Known Activations