INDEX

Explanations

attributions or acknowledgments within a text

oai_token-act-pair · gpt-3.5-turbo

references to authorities or citations in discussions

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

emate

-0.82

rontal

-0.71

istries

-0.69

Tot

-0.66

neutral

-0.65

chrome

-0.62

blem

-0.61

onite

-0.60

rig

-0.59

obal

-0.58

POSITIVE LOGITS

 aptly

0.68

 pointed

0.66

 eloqu

0.65

 optim

0.65

 dictates

0.65

lance

0.64

imum

0.62

 Thrones

0.62

iHUD

0.61

Sov

0.60

Activations Density 0.171%

attributions or acknowledgments within a text

references to authorities or citations in discussions

No Comments

No Known Activations