INDEX

Explanations

mentions of authority figures or entities

oai_token-act-pair · gpt-3.5-turbo

words that emphasize importance or significance

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

reports

-0.70

 Vaughan

-0.70

 Rounds

-0.69

 credits

-0.69

 Emails

-0.68

 reports

-0.67

 upgrades

-0.66

ographs

-0.66

 alerts

-0.65

 Originally

-0.65

POSITIVE LOGITS

rouse

1.15

lot

1.06

usterity

1.04

 bunch

0.96

 finite

0.94

 particular

0.93

versive

0.93

 plethora

0.92

few

0.91

 hypothetical

0.91

Activations Density 0.905%

mentions of authority figures or entities

words that emphasize importance or significance

No Comments

No Known Activations