INDEX

Explanations

numbers at the beginning of phrases

oai_token-act-pair · gpt-3.5-turbo

structured numerical identifiers or listings

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

 tremend

-0.84

ileaks

-0.80

Ô

-0.78

 behavi

-0.78

ifully

-0.77

chwitz

-0.77

anooga

-0.77

ierrez

-0.76

merce

-0.76

icist

-0.75

POSITIVE LOGITS

1.01

0.99

0.96

 Apostles

0.94

0.93

noon

0.92

0.91

Activations Density 0.046%

numbers at the beginning of phrases

structured numerical identifiers or listings

No Comments

No Known Activations