INDEX

Explanations

proper names or titles mentioned in text

oai_token-act-pair · gpt-3.5-turbo

references to names and titles

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

onga

-0.79

ometers

-0.71

ersive

-0.65

oun

-0.64

erker

-0.63

gif

-0.62

aukee

-0.62

eren

-0.62

 contrace

-0.61

umat

-0.61

POSITIVE LOGITS

"#

0.95

''

0.86

"@

0.82

onyms

0.74

"(

0.72

``

0.71

name

0.70

é£

0.70

 Osama

0.70

"_

0.69

Activations Density 0.104%

proper names or titles mentioned in text

references to names and titles

No Comments

No Known Activations