INDEX

Explanations

personal names

oai_token-act-pair · gpt-3.5-turbo

words that possibly indicate engagement or performance in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Reincarn

-0.67

 sovere

-0.64

 Prometheus

-0.63

UID

-0.58

OUS

-0.56

 DATA

-0.55

 Annotations

-0.54

Redd

-0.53

 networking

-0.52

 precursor

-0.52

POSITIVE LOGITS

zinski

1.11

zik

0.97

imir

0.85

hov

0.83

cko

0.82

odore

0.80

isi

0.79

enei

0.79

chy

0.78

itzer

0.75

Activations Density 0.143%

personal names

words that possibly indicate engagement or performance in various contexts

No Comments

No Known Activations