INDEX

Explanations

dates from the 1990s

oai_token-act-pair · gpt-3.5-turbo

references to specific decades in the 20th century

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Þ

-1.04

 exting

-0.98

ô

-0.94

ñ

-0.94

 pione

-0.88

ò

-0.87

 eleph

-0.87

 practition

-0.86

aditional

-0.85

©¶æ

-0.85

POSITIVE LOGITS

1.46

ses

1.13

ties

1.01

ixties

0.96

sburg

0.92

sis

0.90

sen

0.90

Gs

0.86

sci

0.85

zes

0.84

Activations Density 0.054%

dates from the 1990s

references to specific decades in the 20th century

No Comments

No Known Activations