INDEX

Explanations

specific, distinct keywords, likely related to a particular topic or theme

oai_token-act-pair · gpt-3.5-turbo

phrases or concepts related to known entities or characteristics

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

Tokens

-0.64

 helicop

-0.62

 toget

-0.61

enegger

-0.61

 contrace

-0.60

 thous

-0.60

ividually

-0.59

 nodd

-0.58

obbies

-0.57

amac

-0.57

POSITIVE LOGITS

Ĥª

0.62

CRC

0.58

UDP

0.56

 Labour

0.55

 Clash

0.54

SPI

0.54

 OFFIC

0.53

Psy

0.53

 Disclosure

0.53

 SpaceX

0.53

Activations Density 0.578%

specific, distinct keywords, likely related to a particular topic or theme

phrases or concepts related to known entities or characteristics

No Comments

No Known Activations