INDEX

Explanations

names related to various entities or events

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to games and media titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

abases

-0.84

ulates

-0.76

uphem

-0.76

perature

-0.74

nesses

-0.74

yright

-0.74

lying

-0.73

acters

-0.72

ety

-0.71

sheet

-0.71

POSITIVE LOGITS

 Lounge

1.05

 Pictures

1.00

 Entertainment

1.00

 Labs

1.00

 Brewing

0.97

 Technologies

0.96

 Records

0.96

 Studios

0.95

 Rising

0.93

 Farms

0.92

Activations Density 0.252%

names related to various entities or events

proper nouns related to games and media titles

No Comments

No Known Activations