INDEX

Explanations

words related to changes in policies, rules, or regulations

oai_token-act-pair · gpt-3.5-turbo

phrases indicating changes or modifications to existing systems or processes

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

horn

-0.73

IPP

-0.65

Got

-0.63

 Cosponsors

-0.63

 SERV

-0.63

esi

-0.63

leground

-0.63

 PLUS

-0.63

 breaths

-0.61

croft

-0.60

POSITIVE LOGITS

 ensure

0.93

 reduce

0.87

 improve

0.87

 eliminate

0.86

 compensate

0.86

 avoid

0.85

dos

0.85

 overcome

0.84

 clarify

0.84

 postpone

0.83

Activations Density 0.133%

words related to changes in policies, rules, or regulations

phrases indicating changes or modifications to existing systems or processes

No Comments

No Known Activations