INDEX

Explanations

adjectives describing quantity or size

oai_token-act-pair · gpt-3.5-turbo

terms related to large-scale events or issues

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

Ïī

-0.84

present

-0.70

Downloadha

-0.70

ãĤ¼

-0.69

åĤ

-0.68

ãģ¾

-0.66

iphate

-0.65

 thereof

-0.65

slot

-0.65

AX

-0.64

POSITIVE LOGITS

 Examples

1.00

 Thoughts

0.93

ly

0.93

 Problems

0.90

 Locations

0.90

 Techniques

0.86

 Categories

0.86

 Features

0.84

 Names

0.84

 Stories

0.83

Activations Density 0.221%

adjectives describing quantity or size

terms related to large-scale events or issues

No Comments

No Known Activations