INDEX

Explanations

words related to specific actions and events happening in different scenarios

oai_token-act-pair · gpt-3.5-turbo

references to photography and visual imagery

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

partisan

-0.58

 Founders

-0.54

 Tale

-0.51

¿½

-0.51

netflix

-0.49

 Revision

-0.49

milo

-0.48

 Gamble

-0.48

 Patreon

-0.47

encers

-0.46

POSITIVE LOGITS

VERTISEMENT

0.58

ifice

0.55

hesion

0.53

imeters

0.51

 unsuspecting

0.51

 unnoticed

0.50

escape

0.49

 undet

0.49

 guiActiveUn

0.49

 prey

0.49

Activations Density 2.486%

words related to specific actions and events happening in different scenarios

references to photography and visual imagery

No Comments

No Known Activations